Rohan Chandra

Anthony Francis, Claudia Pérez-D'Arpino, Chengshu Li et al. · cmu, mit

A major challenge to deploying robots widely is navigation in human-populated environments, commonly referred to as social robot navigation. While the field of social navigation has advanced tremendously in recent years, the fair evaluation of algorithms that tackle social navigation remains hard because it involves not just robotic agents moving in static environments but also dynamic human agents and their perceptions of the appropriateness of robot behavior. In contrast, clear, repeatable, and accessible benchmarks have accelerated progress in fields like computer vision, natural language processing and traditional robot navigation by enabling researchers to fairly compare algorithms, revealing limitations of existing solutions and illuminating promising new directions. We believe the same approach can benefit social navigation. In this paper, we pave the road towards common, widely accessible, and repeatable benchmarking criteria to evaluate social robot navigation. Our contributions include (a) a definition of a socially navigating robot as one that respects the principles of safety, comfort, legibility, politeness, social competency, agent understanding, proactivity, and responsiveness to context, (b) guidelines for the use of metrics, development of scenarios, benchmarks, datasets, and simulators to evaluate social navigation, and (c) a design of a social navigation metrics framework to make it easier to compare results from different simulators, robots and datasets.

Xiyang Wu, Rohan Chandra, Tianrui Guan et al.

Navigating safely and efficiently in dense and heterogeneous traffic scenarios is challenging for autonomous vehicles (AVs) due to their inability to infer the behaviors or intentions of nearby drivers. In this work, we introduce a distributed multi-agent reinforcement learning (MARL) algorithm that can predict trajectories and intents in dense and heterogeneous traffic scenarios. Our approach for intent-aware planning, iPLAN, allows agents to infer nearby drivers' intents solely from their local observations. We model two distinct incentives for agents' strategies: Behavioral Incentive for high-level decision-making based on their driving behavior or personality and Instant Incentive for motion planning for collision avoidance based on the current traffic state. Our approach enables agents to infer their opponents' behavior incentives and integrate this inferred information into their decision-making and motion-planning processes. We perform experiments on two simulation environments, Non-Cooperative Navigation and Heterogeneous Highway. In Heterogeneous Highway, results show that, compared with centralized training decentralized execution (CTDE) MARL baselines such as QMIX and MAPPO, our method yields a 4.3% and 38.4% higher episodic reward in mild and chaotic traffic, with 48.1% higher success rate and 80.6% longer survival time in chaotic traffic. We also compare with a decentralized training decentralized execution (DTDE) baseline IPPO and demonstrate a higher episodic reward of 12.7% and 6.3% in mild traffic and chaotic traffic, 25.3% higher success rate, and 13.7% longer survival time.

Chen Tang, Ben Abbatematteo, Jiaheng Hu et al.

Reinforcement learning (RL), particularly its combination with deep neural networks referred to as deep RL (DRL), has shown tremendous promise across a wide range of applications, suggesting its potential for enabling the development of sophisticated robotic behaviors. Robotics problems, however, pose fundamental difficulties for the application of RL, stemming from the complexity and cost of interacting with the physical world. This article provides a modern survey of DRL for robotics, with a particular focus on evaluating the real-world successes achieved with DRL in realizing several key robotic competencies. Our analysis aims to identify the key factors underlying those exciting successes, reveal underexplored areas, and provide an overall characterization of the status of DRL in robotics. We highlight several important avenues for future work, emphasizing the need for stable and sample-efficient real-world RL paradigms, holistic approaches for discovering and integrating various competencies to tackle complex long-horizon, open-world tasks, and principled development and evaluation procedures. This survey is designed to offer insights for both RL practitioners and roboticists toward harnessing RL's power to create generally capable real-world robotic systems.

Rohan Chandra, Rahul Maligi, Arya Anantula et al.

We propose an extension to the MAPF formulation, called SocialMAPF, to account for private incentives of agents in constrained environments such as doorways, narrow hallways, and corridor intersections. SocialMAPF is able to, for instance, accurately reason about the urgent incentive of an agent rushing to the hospital over another agent's less urgent incentive of going to a grocery store; MAPF ignores such agent-specific incentives. Our proposed formulation addresses the open problem of optimal and efficient path planning for agents with private incentives. To solve SocialMAPF, we propose a new class of algorithms that use mechanism design during conflict resolution to simultaneously optimize agents' private local utilities and the global system objective. We perform an extensive array of experiments that show that optimal search-based MAPF techniques lead to collisions and increased time-to-goal in SocialMAPF compared to our proposed method using mechanism design. Furthermore, we empirically demonstrate that mechanism design results in models that maximizes agent utility and minimizes the overall time-to-goal of the entire system. We further showcase the capabilities of mechanism design-based planning by successfully deploying it in environments with static obstacles. To conclude, we briefly list several research directions using the SocialMAPF formulation, such as exploring motion planning in the continuous domain for agents with private incentives.

Zayne Sprague, Rohan Chandra, Jarrett Holtz et al.

We present SocialGym 2, a multi-agent navigation simulator for social robot research. Our simulator models multiple autonomous agents, replicating real-world dynamics in complex environments, including doorways, hallways, intersections, and roundabouts. Unlike traditional simulators that concentrate on single robots with basic kinematic constraints in open spaces, SocialGym 2 employs multi-agent reinforcement learning (MARL) to develop optimal navigation policies for multiple robots with diverse, dynamic constraints in complex environments. Built on the PettingZoo MARL library and Stable Baselines3 API, SocialGym 2 offers an accessible python interface that integrates with a navigation stack through ROS messaging. SocialGym 2 can be easily installed and is packaged in a docker container, and it provides the capability to swap and evaluate different MARL algorithms, as well as customize observation and reward functions. We also provide scripts to allow users to create their own environments and have conducted benchmarks using various social navigation algorithms, reporting a broad range of social navigation metrics. Projected hosted at: https://amrl.cs.utexas.edu/social_gym/index.html

Rohan Chandra, Rahul Menon, Zayne Sprague et al.

This paper presents a fully decentralized approach for realtime non-cooperative multi-robot navigation in social mini-games, such as navigating through a narrow doorway or negotiating right of way at a corridor intersection. Our contribution is a new realtime bi-level optimization algorithm, in which the top-level optimization consists of computing a fair and collision-free ordering followed by the bottom-level optimization which plans optimal trajectories conditioned on the ordering. We show that, given such a priority order, we can impose simple kinodynamic constraints on each robot that are sufficient for it to plan collision-free trajectories with minimal deviation from their preferred velocities, similar to how humans navigate in these scenarios. We successfully deploy the proposed algorithm in the real world using F$1/10$ robots, a Clearpath Jackal, and a Boston Dynamics Spot as well as in simulation using the SocialGym 2.0 multi-agent social navigation simulator, in the doorway and corridor intersection scenarios. We compare with state-of-the-art social navigation methods using multi-agent reinforcement learning, collision avoidance algorithms, and crowd simulation models. We show that $(i)$ classical navigation performs $44\%$ better than the state-of-the-art learning-based social navigation algorithms, $(ii)$ without a scheduling protocol, our approach results in collisions in social mini-games $(iii)$ our approach yields $2\times$ and $5\times$ fewer velocity changes than CADRL in doorways and intersections, and finally $(iv)$ bi-level navigation in doorways at a flow rate of $2.8 - 3.3$ (ms)$^{-1}$ is comparable to flow rate in human navigation at a flow rate of $4$ (ms)$^{-1}$.

Rohan Chandra, Shubham Singh, Wenhao Luo et al.

The "Last Mile Challenge" has long been considered an important, yet unsolved, challenge for autonomous vehicles, public service robots, and delivery robots. A central issue in this challenge is the ability of robots to navigate constrained and cluttered environments that have high agency (e.g., doorways, hallways, corridor intersections), often while competing for space with other robots and humans. We refer to these environments as "Social Mini-Games" (SMGs). Traditional navigation approaches designed for MRN do not perform well in SMGs, which has led to focused research on dedicated SMG solvers. However, publications on SMG navigation research make different assumptions, and have different objective functions (safety versus liveness). These assumptions and objectives are sometimes implicitly assumed or described informally. This makes it difficult to establish appropriate baselines for comparison in research papers, as well as making it difficult for practitioners to find the papers relevant to their concrete application. Such ad-hoc representation of the field also presents a barrier to new researchers wanting to start research in this area. SMG navigation research requires its own taxonomy, definitions, and evaluation protocols to guide effective research moving forward. This survey is the first to catalog SMG solvers using a well-defined and unified taxonomy and to classify existing methods accordingly. It also discusses the essential properties of SMG solvers, defines what SMGs are and how they appear in practice, outlines how to evaluate SMG solvers, and highlights the differences between SMG solvers and general navigation systems. The survey concludes with an overview of future directions and open challenges in the field. Our project is open-sourced at https://socialminigames.github.io/{https://socialminigames.github.io/.

Vagul Mahadevan, Shangtong Zhang, Rohan Chandra

Safe, agile, and socially compliant multi-robot navigation in cluttered and constrained environments remains a critical challenge. This is especially difficult with self-interested agents with unique, unknown priorities in decentralized settings, where there is no central authority to resolve conflicts induced by spatial symmetry. We address this challenge by proposing an intuitive, but very effective approach, GameChat, which facilitates safe, agile, and deadlock-free navigation for both cooperative and self-interested agents in cluttered environments. Key to our approach is the idea that agents should resolve conflicts on their own using natural language to communicate, much like humans. We evaluate GameChat in simulated environments with doorways and intersections. The results show that even in the worst case, GameChat reduces the time for all agents to reach their goals by over 35% from a naive baseline and by over 20% from a state of the art baseline in the intersection scenario, while doubling the rate of ensuring the agent with a higher priority task reaches the goal first, from 50% (equivalent to random chance) to 100%. We also demonstrate how GameChat can be extended to more than two agents.

Wentao Zhou, Xuweiyi Chen, Vignesh Rajagopal et al.

The success of foundation models in language and vision motivated research in fully end-to-end robot navigation foundation models (NFMs). NFMs directly map monocular visual input to control actions and ignore mid-level vision modules (tracking, depth estimation, etc) entirely. While the assumption that vision capabilities will emerge implicitly is compelling, it requires large amounts of pixel-to-action supervision that are difficult to obtain. The challenge is especially pronounced in dynamic and unstructured settings, where robust navigation requires precise geometric and dynamic understanding, while the depth-scale ambiguity in monocular views further limits accurate spatial reasoning. In this paper, we show that relying on monocular vision and ignoring mid-level vision priors is inefficient. We present StereoWalker, which augments NFMs with stereo inputs and explicit mid-level vision such as depth estimation and dense pixel tracking. Our intuition is straightforward: stereo inputs resolve the depth-scale ambiguity, and modern mid-level vision models provide reliable geometric and motion structure in dynamic scenes. We also curate a large stereo navigation dataset with automatic action annotation from Internet stereo videos to support training of StereoWalker and to facilitate future research. Through our experiments, we find that mid-level vision enables StereoWalker to achieve a comparable performance as the state-of-the-art using only 1.5% of the training data, and surpasses the state-of-the-art using the full data. We also observe that stereo vision yields higher navigation performance than monocular input.

Zixuan Xie, Xinyu Liu, Claire Chen et al.

In-context reinforcement learning (ICRL) studies agents that, after pretraining, adapt to new tasks by conditioning on additional context without parameter updates. Existing theoretical analyses of ICRL largely rely on linear attention, which replaces the softmax function in the standard attention with an identity mapping. This paper provides the first theoretical understanding of ICRL without making the unrealistic linear attention simplification. In particular, we consider the standard softmax attention used in practice. We show that, with certain parameters, the layerwise forward pass of a Transformer with such softmax attention is equivalent to iterative updates of a weighted softmax temporal difference (TD) learning algorithm. Here, weighted softmax TD is a new RL algorithm that performs policy evaluation in kernel space and adopts both linear TD and tabular TD as special cases. We also prove that under a certain contraction condition, the policy evaluation error decays as the number of layers grows, with the identified parameters above. Finally, we prove that those parameters are a global minimizer of a pretraining loss, explaining their emergence in our numerical experiments.

Tianrui Guan, Jun Wang, Shiyi Lan et al.

We present a novel architecture for 3D object detection, M3DeTR, which combines different point cloud representations (raw, voxels, bird-eye view) with different feature scales based on multi-scale feature pyramids. M3DeTR is the first approach that unifies multiple point cloud representations, feature scales, as well as models mutual relationships between point clouds simultaneously using transformers. We perform extensive ablation experiments that highlight the benefits of fusing representation and scale, and modeling the relationships. Our method achieves state-of-the-art performance on the KITTI 3D object detection dataset and Waymo Open Dataset. Results show that M3DeTR improves the baseline significantly by 1.48% mAP for all classes on Waymo Open Dataset. In particular, our approach ranks 1st on the well-known KITTI 3D Detection Benchmark for both car and cyclist classes, and ranks 1st on Waymo Open Dataset with single frame point cloud input. Our code is available at: https://github.com/rayguan97/M3DETR.

Zixuan Xie, Xinyu Liu, Rohan Chandra et al.

In-context reinforcement learning (ICRL) refers to the ability of RL agents to adapt to new tasks at inference time without parameter updates by conditioning on additional context. Recent empirical studies further demonstrate that Chain-of-Thought (CoT) generation can amplify this ICRL capability. This paper is the first to provide a theoretical understanding on how CoT interacts with ICRL. We conduct our analysis in a policy evaluation setup with linear Transformer. We prove that with specific Transformer parameters, the CoT generation process is equivalent to repeatedly executing temporal difference learning updates. Additionally, we provide finite sample convergence analysis showing that the policy evaluation error decreases geometrically with CoT length and eventually saturates at a statistical floor determined by the context length. We also prove that the desired Transformer parameters are a global minimizer of the pretraining loss, providing a theoretical understanding on the empirical emergence of those parameters.

Amir Moeini, Jiuqi Wang, Jacob Beck et al.

Reinforcement learning (RL) agents typically optimize their policies by performing expensive backward passes to update their network parameters. However, some agents can solve new tasks without updating any parameters by simply conditioning on additional context such as their action-observation histories. This paper surveys work on such behavior, known as in-context reinforcement learning.

Kefan Song, Amir Moeini, Peng Wang et al.

Reinforcement learning (RL) is a human-designed framework for solving sequential decision making problems. In this work, we demonstrate that, surprisingly, RL emerges in LLM's (Large Language Model) inference time -- a phenomenon known as in-context RL (ICRL). Specifically, we propose a novel multi-round prompting framework called ICRL prompting. The goal is to prompt the LLM to complete a task. After the LLM generates a response at the current round, we give numerical scalar feedbacks for the response, called the rewards. At the next round, we prompt the LLM again with the same task and a context consisting of all previous responses and rewards. We observe that the quality of the LLM's response increases as the context grows. In other words, the LLM is able to maximize the scalar reward signal in the inference time, just like an RL algorithm. We evaluate ICRL prompting in three benchmarks (Game of 24, creative writing, and ScienceWorld) and demonstrate significant performance improvements over baseline methods such as Self-Refine and Reflexion. Surprisingly, in some experiments the reward signals are generated by the LLM itself, yet performance improvements are still observed from ICRL prompting, offering a promising paradigm for scaling test-time compute.

Jeffrey Chen, Rohan Chandra

Robots operating in dynamic, unstructured environments must balance safety and efficiency under potentially limited sensing. While control barrier functions (CBFs) provide principled collision avoidance via safety filtering, their behavior is often governed by fixed parameters that can be overly conservative in benign scenes or overly permissive near hazards. We present AlphaAdj, a vision-to-control navigation framework that uses egocentric RGB input to adapt the conservativeness of a CBF safety filter in real time. A vision-language model(VLM) produces a bounded scalar risk estimate from the current camera view, which we map to dynamically update a CBF parameter that modulates how strongly safety constraints are enforced. To address asynchronous inference and non-trivial VLM latency in practice, we combine a geometric, speed-aware dynamic cap and a staleness-gated fusion policy with lightweight implementation choices that reduce end-to-end inference overhead. We evaluate AlphaAdj across multiple static and dynamic obstacle scenarios in a variety of environments, comparing against fixed-parameter and uncapped ablations. Results show that AlphaAdj maintains collision-free navigation while improving efficiency (in terms of path length and time to goal) by up to 18.5% relative to fixed settings and improving robustness and success rate relative to an uncapped baseline.

Chirag Parikh, Ravi Shankar Mishra, Rohan Chandra et al.

Recognizing driving behaviors is important for downstream tasks such as reasoning, planning, and navigation. Existing video recognition approaches work well for common behaviors (e.g. "drive straight", "brake", "turn left/right"). However, the performance is sub-par for underrepresented/rare behaviors typically found in tail of the behavior class distribution. To address this shortcoming, we propose Transfer-LMR, a modular training routine for improving the recognition performance across all driving behavior classes. We extensively evaluate our approach on METEOR and HDD datasets that contain rich yet heavy-tailed distribution of driving behaviors and span diverse traffic scenarios. The experimental results demonstrate the efficacy of our approach, especially for recognizing underrepresented/rare driving behaviors.

Jiuqi Wang, Rohan Chandra, Shangtong Zhang

Loss of plasticity is one of the main challenges in continual learning with deep neural networks, where neural networks trained via backpropagation gradually lose their ability to adapt to new tasks and perform significantly worse than their freshly initialized counterparts. The main contribution of this paper is to propose a new hypothesis that experience replay addresses the loss of plasticity in continual learning. Here, experience replay is a form of memory. We provide supporting evidence for this hypothesis. In particular, we demonstrate in multiple different tasks, including regression, classification, and policy evaluation, that by simply adding an experience replay and processing the data in the experience replay with Transformers, the loss of plasticity disappears. Notably, we do not alter any standard components of deep learning. For example, we do not change backpropagation. We do not modify the activation functions. And we do not use any regularization. We conjecture that experience replay and Transformers can address the loss of plasticity because of the in-context learning phenomenon.

Amir Moeini, Minjae Kwon, Alper Kamil Bozkurt et al.

In-context reinforcement learning (ICRL) is an emerging RL paradigm where the agent, after some pretraining procedure, is able to adapt to out-of-distribution test tasks without any parameter updates. The agent achieves this by continually expanding the input (i.e., the context) to its policy neural networks. For example, the input could be all the history experience that the agent has access to until the current time step. The agent's performance improves as the input grows, without any parameter updates. In this work, we propose the first method that promotes the safety of ICRL's adaptation process in the framework of constrained Markov Decision Processes. In other words, during the parameter-update-free adaptation process, the agent not only maximizes the reward but also minimizes an additional cost function. We also demonstrate that our agent actively reacts to the threshold (i.e., budget) of the cost tolerance. With a higher cost budget, the agent behaves more aggressively, and with a lower cost budget, the agent behaves more conservatively.

Jiuqi Wang, Rohan Chandra, Shangtong Zhang

Typically, a modern reinforcement learning (RL) agent solves a task by updating its neural network parameters to adapt its policy to the task. Recently, it has been observed that some RL agents can solve a wide range of new out-of-distribution tasks without parameter updates after pretraining on some task distribution. When evaluated in a new task, instead of making parameter updates, the pretrained agent conditions its policy on additional input called the context, e.g., the agent's interaction history in the new task. The agent's performance increases as the information in the context increases, with the agent's parameters fixed. This phenomenon is typically called in-context RL (ICRL). The pretrained parameters of the agent network enable the remarkable ICRL phenomenon. However, many ICRL works perform the pretraining with standard RL algorithms. This raises the central question this paper aims to address: Why can the RL pretraining algorithm generate network parameters that enable ICRL? We hypothesize that the parameters capable of ICRL are minimizers of the pretraining loss. This work provides initial support for this hypothesis through a case study. In particular, we prove that when a Transformer is pretrained for policy evaluation, one of the global minimizers of the pretraining loss can enable in-context temporal difference learning.

Zixuan Xie, Xinyu Liu, Rohan Chandra et al.

Linear TD($λ$) is one of the most fundamental reinforcement learning algorithms for policy evaluation. Previously, convergence rates are typically established under the assumption of linearly independent features, which does not hold in many practical scenarios. This paper instead establishes the first $L^2$ convergence rates for linear TD($λ$) operating under arbitrary features, without making any algorithmic modification or additional assumptions. Our results apply to both the discounted and average-reward settings. To address the potential non-uniqueness of solutions resulting from arbitrary features, we develop a novel stochastic approximation result featuring convergence rates to the solution set instead of a single point.

Aleesha Khurram, Amir Moeini, Shangtong Zhang et al.

Despite significant progress and advances in autonomous driving, many end-to-end systems still struggle with domain adaptation (DA), such as transferring a policy trained under clear weather to adverse weather conditions. Typical DA strategies in the literature include collecting additional data in the target domain or re-training the model, or both. Both these strategies quickly become impractical as we increase scale and complexity of driving. These limitations have encouraged investigation into few-shot and zero-shot prompt-driven DA at inference time involving LLMs and VLMs. These methods work by adding a few state-action trajectories during inference to the prompt (similar to in-context learning). However, there are two limitations of such an approach: $(i)$ prompt-driven DA methods are currently restricted to perception tasks such as detection and segmentation and $(ii)$ they require expert few-shot data. In this work, we present a new approach to inference-time few-shot prompt-driven DA for closed-loop autonomous driving in adverse weather condition using in-context reinforcement learning (ICRL). Similar to other prompt-driven DA methods, our approach does not require any updates to the model parameters nor does it require additional data collection in adversarial weather regime. Furthermore, our approach advances the state-of-the-art in prompt-driven DA by extending to closed driving using general trajectories observed during inference. Our experiments using the CARLA simulator show that ICRL results in safer, more efficient, and more comfortable driving policies in the target domain compared to state-of-the-art prompt-driven DA baselines.

Timur Anvar, Jeffrey Chen, Yuyan Wang et al.

Autonomous vehicle navigation in complex environments such as dense and fast-moving highways and merging scenarios remains an active area of research. A key limitation of RL is its reliance on well-specified reward functions, which often fail to capture the full semantic and social complexity of diverse, out-of-distribution situations. As a result, a rapidly growing line of research explores using Large Language Models (LLMs) to replace or supplement RL for direct planning and control, on account of their ability to reason about rich semantic context. However, LLMs present significant drawbacks: they can be unstable in zero-shot safety-critical settings, produce inconsistent outputs, and often depend on expensive API calls with network latency. This motivates our investigation into whether small, locally deployed LLMs (< 14B parameters) can meaningfully support autonomous highway driving through reward shaping rather than direct control. We present a case study comparing RL-only, LLM-only, and hybrid approaches, where LLMs augment RL rewards by scoring state-action transitions during training, while standard RL policies execute at test time. Our findings reveal that RL-only agents achieve moderate success rates (73-89%) with reasonable efficiency, LLM-only agents can reach higher success rates (up to 94%) but with severely degraded speed performance, and hybrid approaches consistently fall between these extremes. Critically, despite explicit efficiency instructions, LLM-influenced approaches exhibit systematic conservative bias with substantial model-dependent variability, highlighting important limitations of current small LLMs for safety-critical control tasks.

Kefan Song, Runnan Jiang, Rohan Chandra et al.

This paper addresses a critical societal consideration in the application of Reinforcement Learning (RL): ensuring equitable outcomes across different demographic groups in multi-task settings. While previous work has explored fairness in single-task RL, many real-world applications are multi-task in nature and require policies to maintain fairness across all tasks. We introduce a novel formulation of multi-task group fairness in RL and propose a constrained optimization algorithm that explicitly enforces fairness constraints across multiple tasks simultaneously. We have shown that our proposed algorithm does not violate fairness constraints with high probability and with sublinear regret in the finite-horizon episodic setting. Through experiments in RiverSwim and MuJoCo environments, we demonstrate that our approach better ensures group fairness across multiple tasks compared to previous methods that lack explicit multi-task fairness constraints in both the finite-horizon setting and the infinite-horizon setting. Our results show that the proposed algorithm achieves smaller fairness gaps while maintaining comparable returns across different demographic groups and tasks, suggesting its potential for addressing fairness concerns in real-world multi-task RL applications.

Kefan Song, Jin Yao, Runnan Jiang et al.

As Large Language Models (LLMs) become increasingly powerful and accessible to human users, ensuring fairness across diverse demographic groups, i.e., group fairness, is a critical ethical concern. However, current fairness and bias research in LLMs is limited in two aspects. First, compared to traditional group fairness in machine learning classification, it requires that the non-sensitive attributes, in this case, the prompt questions, be the same across different groups. In many practical scenarios, different groups, however, may prefer different prompt questions and this requirement becomes impractical. Second, it evaluates group fairness only for the LLM's final output without identifying the source of possible bias. Namely, the bias in LLM's output can result from both the pretraining and the finetuning. For finetuning, the bias can result from both the RLHF procedure and the learned reward model. Arguably, evaluating the group fairness of each component in the LLM pipeline could help develop better methods to mitigate the possible bias. Recognizing those two limitations, this work benchmarks the group fairness of learned reward models. By using expert-written text from arXiv, we are able to benchmark the group fairness of reward models without requiring the same prompt questions across different demographic groups. Surprisingly, our results demonstrate that all the evaluated reward models (e.g., Nemotron-4-340B-Reward, ArmoRM-Llama3-8B-v0.1, and GRM-llama3-8B-sftreg) exhibit statistically significant group unfairness. We also observed that top-performing reward models (w.r.t. canonical performance metrics) tend to demonstrate better group fairness.

Xijun Wang, Pedro Sandoval-Segura, Chengyuan Zhang et al.

Most existing traffic video datasets including Waymo are structured, focusing predominantly on Western traffic, which hinders global applicability. Specifically, most Asian scenarios are far more complex, involving numerous objects with distinct motions and behaviors. Addressing this gap, we present a new dataset, DAVE, designed for evaluating perception methods with high representation of Vulnerable Road Users (VRUs: e.g. pedestrians, animals, motorbikes, and bicycles) in complex and unpredictable environments. DAVE is a manually annotated dataset encompassing 16 diverse actor categories (spanning animals, humans, vehicles, etc.) and 16 action types (complex and rare cases like cut-ins, zigzag movement, U-turn, etc.), which require high reasoning ability. DAVE densely annotates over 13 million bounding boxes (bboxes) actors with identification, and more than 1.6 million boxes are annotated with both actor identification and action/behavior details. The videos within DAVE are collected based on a broad spectrum of factors, such as weather conditions, the time of day, road scenarios, and traffic density. DAVE can benchmark video tasks like Tracking, Detection, Spatiotemporal Action Localization, Language-Visual Moment retrieval, and Multi-label Video Action Recognition. Given the critical importance of accurately identifying VRUs to prevent accidents and ensure road safety, in DAVE, vulnerable road users constitute 41.13% of instances, compared to 23.71% in Waymo. DAVE provides an invaluable resource for the development of more sensitive and accurate visual perception algorithms in the complex real world. Our experiments show that existing methods suffer degradation in performance when evaluated on DAVE, highlighting its benefit for future video recognition research.

Nilesh Suriyarachchi, Rohan Chandra, John S. Baras et al.

We propose GameOpt: a novel hybrid approach to cooperative intersection control for dynamic, multi-lane, unsignalized intersections. Safely navigating these complex and accident prone intersections requires simultaneous trajectory planning and negotiation among drivers. GameOpt is a hybrid formulation that first uses an auction mechanism to generate a priority entrance sequence for every agent, followed by an optimization-based trajectory planner that computes velocity controls that satisfy the priority sequence. This coupling operates at real-time speeds of less than 10 milliseconds in high density traffic of more than 10,000 vehicles/hr, 100 times faster than other fully optimization-based methods, while providing guarantees in terms of fairness, safety, and efficiency. Tested on the SUMO simulator, our algorithm improves throughput by at least 25%, time taken to reach the goal by 75%, and fuel consumption by 33% compared to auction-based approaches and signaled approaches using traffic-lights and stop signs.

Rohan Chandra, Aniket Bera, Dinesh Manocha

Studies have shown that autonomous vehicles (AVs) behave conservatively in a traffic environment composed of human drivers and do not adapt to local conditions and socio-cultural norms. It is known that socially aware AVs can be designed if there exists a mechanism to understand the behaviors of human drivers. We present an approach that leverages machine learning to predict, the behaviors of human drivers. This is similar to how humans implicitly interpret the behaviors of drivers on the road, by only observing the trajectories of their vehicles. We use graph-theoretic tools to extract driver behavior features from the trajectories and machine learning to obtain a computational mapping between the extracted trajectory of a vehicle in traffic and the driver behaviors. Compared to prior approaches in this domain, we prove that our method is robust, general, and extendable to broad-ranging applications such as autonomous navigation. We evaluate our approach on real-world traffic datasets captured in the U.S., India, China, and Singapore, as well as in simulation.

Rohan Chandra, Xijun Wang, Mridul Mahajan et al.

We present a new traffic dataset, METEOR, which captures traffic patterns and multi-agent driving behaviors in unstructured scenarios. METEOR consists of more than 1000 one-minute videos, over 2 million annotated frames with bounding boxes and GPS trajectories for 16 unique agent categories, and more than 13 million bounding boxes for traffic agents. METEOR is a dataset for rare and interesting, multi-agent driving behaviors that are grouped into traffic violations, atypical interactions, and diverse scenarios. Every video in METEOR is tagged using a diverse range of factors corresponding to weather, time of the day, road conditions, and traffic density. We use METEOR to benchmark perception methods for object detection and multi-agent behavior prediction. Our key finding is that state-of-the-art models for object detection and behavior prediction, which otherwise succeed on existing datasets such as Waymo, fail on the METEOR dataset. METEOR marks the first step towards the development of more sophisticated perception models for dense, heterogeneous, and unstructured scenarios.

Rohan Chandra, Dinesh Manocha

We present a new method for multi-agent planning involving human drivers and autonomous vehicles (AVs) in unsignaled intersections, roundabouts, and during merging. In multi-agent planning, the main challenge is to predict the actions of other agents, especially human drivers, as their intentions are hidden from other agents. Our algorithm uses game theory to develop a new auction, called GamePlan, that directly determines the optimal action for each agent based on their driving style (which is observable via commonly available sensors). GamePlan assigns a higher priority to more aggressive or impatient drivers and a lower priority to more conservative or patient drivers; we theoretically prove that such an approach is game-theoretically optimal prevents collisions and deadlocks. We compare our approach with prior state-of-the-art auction techniques including economic auctions, time-based auctions (first-in first-out), and random bidding and show that each of these methods result in collisions among agents when taking into account driver behavior. We compare with methods based on DRL, deep learning, and game theory and present our benefits over these approaches. Finally, we show that our approach can be implemented in the real-world with human drivers.

Tianrui Guan, Divya Kothandaraman, Rohan Chandra et al.

We propose GANav, a novel group-wise attention mechanism to identify safe and navigable regions in off-road terrains and unstructured environments from RGB images. Our approach classifies terrains based on their navigability levels using coarse-grained semantic segmentation. Our novel group-wise attention loss enables any backbone network to explicitly focus on the different groups' features with low spatial resolution. Our design leads to efficient inference while maintaining a high level of accuracy compared to existing SOTA methods. Our extensive evaluations on the RUGD and RELLIS-3D datasets shows that GANav achieves an improvement over the SOTA mIoU by 2.25-39.05% on RUGD and 5.17-19.06% on RELLIS-3D. We interface GANav with a deep reinforcement learning-based navigation algorithm and highlight its benefits in terms of navigation in real-world unstructured terrains. We integrate our GANav-based navigation algorithm with ClearPath Jackal and Husky robots, and observe an increase of 10% in terms of success rate, 2-47% in terms of selecting the surface with the best navigability and a decrease of 4.6-13.9% in trajectory roughness. Further, GANav reduces the false positive rate of forbidden regions by 37.79%. Code, videos, and a full technical report are available at https://gamma.umd.edu/offroad/.

Divya Kothandaraman, Rohan Chandra, Dinesh Manocha

We present a novel approach for unsupervised road segmentation in adverse weather conditions such as rain or fog. This includes a new algorithm for source-free domain adaptation (SFDA) using self-supervised learning. Moreover, our approach uses several techniques to address various challenges in SFDA and improve performance, including online generation of pseudo-labels and self-attention as well as use of curriculum learning, entropy minimization and model distillation. We have evaluated the performance on $6$ datasets corresponding to real and synthetic adverse weather conditions. Our method outperforms all prior works on unsupervised road segmentation and SFDA by at least 10.26%, and improves the training time by 18-180x. Moreover, our self-supervised algorithm exhibits similar accuracy performance in terms of mIOU score as compared to prior supervised methods.

Rohan Chandra, Aniket Bera, Dinesh Manocha

Studies have shown that autonomous vehicles (AVs) behave conservatively in a traffic environment composed of human drivers and do not adapt to local conditions and socio-cultural norms. It is known that socially aware AVs can be designed if there exist a mechanism to understand the behaviors of human drivers. We present a notion of Machine Theory of Mind (M-ToM) to infer the behaviors of human drivers by observing the trajectory of their vehicles. Our M-ToM approach, called StylePredict, is based on trajectory analysis of vehicles, which has been investigated in robotics and computer vision. StylePredict mimics human ToM to infer driver behaviors, or styles, using a computational mapping between the extracted trajectory of a vehicle in traffic and the driver behaviors using graph-theoretic techniques, including spectral analysis and centrality functions. We use StylePredict to analyze driver behavior in different cultures in the USA, China, India, and Singapore, based on traffic density, heterogeneity, and conformity to traffic rules and observe an inverse correlation between longitudinal (overspeeding) and lateral (overtaking, lane-changes) driving styles.

Angelos Mavrogiannis, Rohan Chandra, Dinesh Manocha

We address the problem of ego-vehicle navigation in dense simulated traffic environments populated by road agents with varying driver behaviors. Navigation in such environments is challenging due to unpredictability in agents' actions caused by their heterogeneous behaviors. We present a new simulation technique consisting of enriching existing traffic simulators with behavior-rich trajectories corresponding to varying levels of aggressiveness. We generate these trajectories with the help of a driver behavior modeling algorithm. We then use the enriched simulator to train a deep reinforcement learning (DRL) policy that consists of a set of high-level vehicle control commands and use this policy at test time to perform local navigation in dense traffic. Our policy implicitly models the interactions between traffic agents and computes safe trajectories for the ego-vehicle accounting for aggressive driver maneuvers such as overtaking, over-speeding, weaving, and sudden lane changes. Our enhanced behavior-rich simulator can be used for generating datasets that consist of trajectories corresponding to diverse driver behaviors and traffic densities, and our behavior-based navigation scheme can be combined with state-of-the-art navigation algorithms.

Divya Kothandaraman, Rohan Chandra, Dinesh Manocha

We present an unsupervised adaptation approach for visual scene understanding in unstructured traffic environments. Our method is designed for unstructured real-world scenarios with dense and heterogeneous traffic consisting of cars, trucks, two-and three-wheelers, and pedestrians. We describe a new semantic segmentation technique based on unsupervised domain adaptation (DA), that can identify the class or category of each region in RGB images or videos. We also present a novel self-training algorithm (Alt-Inc) for multi-source DA that improves the accuracy. Our overall approach is a deep learning-based technique and consists of an unsupervised neural network that achieves 87.18% accuracy on the challenging India Driving Dataset. Our method works well on roads that may not be well-marked or may include dirt, unidentifiable debris, potholes, etc. A key aspect of our approach is that it can also identify objects that are encountered by the model for the fist time during the testing phase. We compare our method against the state-of-the-art methods and show an improvement of 5.17% - 42.9%. Furthermore, we also conduct user studies that qualitatively validate the improvements in visual scene understanding of unstructured driving environments.

Trisha Mittal, Uttaran Bhattacharya, Rohan Chandra et al.

We present a learning-based method for detecting real and fake deepfake multimedia content. To maximize information for learning, we extract and analyze the similarity between the two audio and visual modalities from within the same video. Additionally, we extract and compare affective cues corresponding to perceived emotion from the two modalities within a video to infer whether the input video is "real" or "fake". We propose a deep learning network, inspired by the Siamese network architecture and the triplet loss. To validate our model, we report the AUC metric on two large-scale deepfake detection datasets, DeepFake-TIMIT Dataset and DFDC. We compare our approach with several SOTA deepfake detection methods and report per-video AUC of 84.4% on the DFDC and 96.6% on the DF-TIMIT datasets, respectively. To the best of our knowledge, ours is the first approach that simultaneously exploits audio and video modalities and also perceived emotions from the two modalities for deepfake detection.

Trisha Mittal, Pooja Guhan, Uttaran Bhattacharya et al.

We present EmotiCon, a learning-based algorithm for context-aware perceived human emotion recognition from videos and images. Motivated by Frege's Context Principle from psychology, our approach combines three interpretations of context for emotion recognition. Our first interpretation is based on using multiple modalities(e.g. faces and gaits) for emotion recognition. For the second interpretation, we gather semantic context from the input image and use a self-attention-based CNN to encode this information. Finally, we use depth maps to model the third interpretation related to socio-dynamic interactions and proximity among agents. We demonstrate the efficiency of our network through experiments on EMOTIC, a benchmark dataset. We report an Average Precision (AP) score of 35.48 across 26 classes, which is an improvement of 7-8 over prior methods. We also introduce a new dataset, GroupWalk, which is a collection of videos captured in multiple real-world settings of people walking. We report an AP of 65.83 across 4 categories on GroupWalk, which is also an improvement over prior methods.

Rohan Chandra, Uttaran Bhattacharya, Trisha Mittal et al.

We present a new measure, CMetric, to classify driver behaviors using centrality functions. Our formulation combines concepts from computational graph theory and social traffic psychology to quantify and classify the behavior of human drivers. CMetric is used to compute the probability of a vehicle executing a driving style, as well as the intensity used to execute the style. Our approach is designed for realtime autonomous driving applications, where the trajectory of each vehicle or road-agent is extracted from a video. We compute a dynamic geometric graph (DGG) based on the positions and proximity of the road-agents and centrality functions corresponding to closeness and degree. These functions are used to compute the CMetric based on style likelihood and style intensity estimates. Our approach is general and makes no assumption about traffic density, heterogeneity, or how driving behaviors change over time. We present an algorithm to compute CMetric and demonstrate its performance on real-world traffic datasets. To test the accuracy of CMetric, we introduce a new evaluation protocol (called "Time Deviation Error") that measures the difference between human prediction and the prediction made by CMetric.

Adarsh Jagan Sathyamoorthy, Jing Liang, Utsav Patel et al.

We present DenseCAvoid, a novel navigation algorithm for navigating a robot through dense crowds and avoiding collisions by anticipating pedestrian behaviors. Our formulation uses visual sensors and a pedestrian trajectory prediction algorithm to track pedestrians in a set of input frames and provide bounding boxes that extrapolate the pedestrian positions in a future time. Our hybrid approach combines this trajectory prediction with a Deep Reinforcement Learning-based collision avoidance method to train a policy to generate smoother, safer, and more robust trajectories during run-time. We train our policy in realistic 3-D simulations of static and dynamic scenarios with multiple pedestrians. In practice, our hybrid approach generalizes well to unseen, real-world scenarios and can navigate a robot through dense crowds (~1-2 humans per square meter) in indoor scenarios, including narrow corridors and lobbies. As compared to cases where prediction was not used, we observe that our method reduces the occurrence of the robot freezing in a crowd by up to 48%, and performs comparably with respect to trajectory lengths and mean arrival times to goal.

Rohan Chandra, Tianrui Guan, Srujan Panuganti et al.

We present a novel approach for traffic forecasting in urban traffic scenarios using a combination of spectral graph analysis and deep learning. We predict both the low-level information (future trajectories) as well as the high-level information (road-agent behavior) from the extracted trajectory of each road-agent. Our formulation represents the proximity between the road agents using a weighted dynamic geometric graph (DGG). We use a two-stream graph-LSTM network to perform traffic forecasting using these weighted DGGs. The first stream predicts the spatial coordinates of road-agents, while the second stream predicts whether a road-agent is going to exhibit overspeeding, underspeeding, or neutral behavior by modeling spatial interactions between road-agents. Additionally, we propose a new regularization algorithm based on spectral clustering to reduce the error margin in long-term prediction (3-5 seconds) and improve the accuracy of the predicted trajectories. Moreover, we prove a theoretical upper bound on the regularized prediction error. We evaluate our approach on the Argoverse, Lyft, Apolloscape, and NGSIM datasets and highlight the benefits over prior trajectory prediction methods. In practice, our approach reduces the average prediction error by approximately 75% over prior algorithms and achieves a weighted average accuracy of 91.2% for behavior prediction. Additionally, our spectral regularization improves long-term prediction by up to 70%.

Uttaran Bhattacharya, Christian Roncal, Trisha Mittal et al.

We present an autoencoder-based semi-supervised approach to classify perceived human emotions from walking styles obtained from videos or motion-captured data and represented as sequences of 3D poses. Given the motion on each joint in the pose at each time step extracted from 3D pose sequences, we hierarchically pool these joint motions in a bottom-up manner in the encoder, following the kinematic chains in the human body. We also constrain the latent embeddings of the encoder to contain the space of psychologically-motivated affective features underlying the gaits. We train the decoder to reconstruct the motions per joint per time step in a top-down manner from the latent embeddings. For the annotated data, we also train a classifier to map the latent embeddings to emotion labels. Our semi-supervised approach achieves a mean average precision of 0.84 on the Emotion-Gait benchmark dataset, which contains both labeled and unlabeled gaits collected from multiple sources. We outperform current state-of-art algorithms for both emotion recognition and action recognition from 3D gaits by 7%--23% on the absolute. More importantly, we improve the average precision by 10%--50% on the absolute on classes that each makes up less than 25% of the labeled part of the Emotion-Gait benchmark dataset.

Trisha Mittal, Uttaran Bhattacharya, Rohan Chandra et al.

We present M3ER, a learning-based method for emotion recognition from multiple input modalities. Our approach combines cues from multiple co-occurring modalities (such as face, text, and speech) and also is more robust than other methods to sensor noise in any of the individual modalities. M3ER models a novel, data-driven multiplicative fusion method to combine the modalities, which learn to emphasize the more reliable cues and suppress others on a per-sample basis. By introducing a check step which uses Canonical Correlational Analysis to differentiate between ineffective and effective modalities, M3ER is robust to sensor noise. M3ER also generates proxy features in place of the ineffectual modalities. We demonstrate the efficiency of our network through experimentation on two benchmark datasets, IEMOCAP and CMU-MOSEI. We report a mean accuracy of 82.7% on IEMOCAP and 89.0% on CMU-MOSEI, which, collectively, is an improvement of about 5% over prior work.

Uttaran Bhattacharya, Trisha Mittal, Rohan Chandra et al.

We present a novel classifier network called STEP, to classify perceived human emotion from gaits, based on a Spatial Temporal Graph Convolutional Network (ST-GCN) architecture. Given an RGB video of an individual walking, our formulation implicitly exploits the gait features to classify the emotional state of the human into one of four emotions: happy, sad, angry, or neutral. We use hundreds of annotated real-world gait videos and augment them with thousands of annotated synthetic gaits generated using a novel generative network called STEP-Gen, built on an ST-GCN based Conditional Variational Autoencoder (CVAE). We incorporate a novel push-pull regularization loss in the CVAE formulation of STEP-Gen to generate realistic gaits and improve the classification accuracy of STEP. We also release a novel dataset (E-Gait), which consists of $2,177$ human gaits annotated with perceived emotions along with thousands of synthetic gaits. In practice, STEP can learn the affective features and exhibits classification accuracy of 89% on E-Gait, which is 14 - 30% more accurate over prior methods.

Rohan Chandra, Uttaran Bhattacharya, Trisha Mittal et al.

We present a novel algorithm (GraphRQI) to identify driver behaviors from road-agent trajectories. Our approach assumes that the road-agents exhibit a range of driving traits, such as aggressive or conservative driving. Moreover, these traits affect the trajectories of nearby road-agents as well as the interactions between road-agents. We represent these inter-agent interactions using unweighted and undirected traffic graphs. Our algorithm classifies the driver behavior using a supervised learning algorithm by reducing the computation to the spectral analysis of the traffic graph. Moreover, we present a novel eigenvalue algorithm to compute the spectrum efficiently. We provide theoretical guarantees for the running time complexity of our eigenvalue algorithm and show that it is faster than previous methods by 2 times. We evaluate the classification accuracy of our approach on traffic videos and autonomous driving datasets corresponding to urban traffic. In practice, GraphRQI achieves an accuracy improvement of up to 25% over prior driver behavior classification algorithms. We also use our classification algorithm to predict the future trajectories of road-agents.

Rohan Chandra, Uttaran Bhattacharya, Christian Roncal et al.

We present RobustTP, an end-to-end algorithm for predicting future trajectories of road-agents in dense traffic with noisy sensor input trajectories obtained from RGB cameras (either static or moving) through a tracking algorithm. In this case, we consider noise as the deviation from the ground truth trajectory. The amount of noise depends on the accuracy of the tracking algorithm. Our approach is designed for dense heterogeneous traffic, where the road agents corresponding to a mixture of buses, cars, scooters, bicycles, or pedestrians. RobustTP is an approach that first computes trajectories using a combination of a non-linear motion model and a deep learning-based instance segmentation algorithm. Next, these noisy trajectories are trained using an LSTM-CNN neural network architecture that models the interactions between road-agents in dense and heterogeneous traffic. Our trajectory prediction algorithm outperforms state-of-the-art methods for end-to-end trajectory prediction using sensor inputs. We achieve an improvement of upto 18% in average displacement error and an improvement ofup to 35.5% in final displacement error at the end of the prediction window (5 seconds) over the next best method. All experiments were set up on an Nvidia TiTan Xp GPU. Additionally, we release a software framework, TrackNPred. The framework consists of implementations of state-of-the-art tracking and trajectory prediction methods and tools to benchmark and evaluate them on real-world dense traffic datasets.

Rohan Chandra, Uttaran Bhattacharya, Tanmay Randhavane et al.

We present a realtime tracking algorithm, RoadTrack, to track heterogeneous road-agents in dense traffic videos. Our approach is designed for traffic scenarios that consist of different road-agents such as pedestrians, two-wheelers, cars, buses, etc. sharing the road. We use the tracking-by-detection approach where we track a road-agent by matching the appearance or bounding box region in the current frame with the predicted bounding box region propagated from the previous frame. RoadTrack uses a novel motion model called the Simultaneous Collision Avoidance and Interaction (SimCAI) model to predict the motion of road-agents by modeling collision avoidance and interactions between the road-agents for the next frame. We demonstrate the advantage of RoadTrack on a dataset of dense traffic videos and observe an accuracy of 75.8% on this dataset, outperforming prior state-of-the-art tracking algorithms by at least 5.2%. RoadTrack operates in realtime at approximately 30 fps and is at least 4 times faster than prior tracking algorithms on standard tracking datasets.

Rohan Chandra, Uttaran Bhattacharya, Aniket Bera et al.

We present a pedestrian tracking algorithm, DensePeds, that tracks individuals in highly dense crowds (greater than 2 pedestrians per square meter). Our approach is designed for videos captured from front-facing or elevated cameras. We present a new motion model called Front-RVO (FRVO) for predicting pedestrian movements in dense situations using collision avoidance constraints and combine it with state-of-the-art Mask R-CNN to compute sparse feature vectors that reduce the loss of pedestrian tracks (false negatives). We evaluate DensePeds on the standard MOT benchmarks as well as a new dense crowd dataset. In practice, our approach is 4.5 times faster than prior tracking algorithms on the MOT benchmark and we are state-of-the-art in dense crowd videos by over 2.6% on the absolute scale on average.

Rohan Chandra, Uttaran Bhattacharya, Aniket Bera et al.

We present a new algorithm for predicting the near-term trajectories of road-agents in dense traffic videos. Our approach is designed for heterogeneous traffic, where the road-agents may correspond to buses, cars, scooters, bicycles, or pedestrians. We model the interactions between different road-agents using a novel LSTM-CNN hybrid network for trajectory prediction. In particular, we take into account heterogeneous interactions that implicitly accounts for the varying shapes, dynamics, and behaviors of different road agents. In addition, we model horizon-based interactions which are used to implicitly model the driving behavior of each road-agent. We evaluate the performance of our prediction algorithm, TraPHic, on the standard datasets and also introduce a new dense, heterogeneous traffic dataset corresponding to urban Asian videos and agent trajectories. We outperform state-of-the-art methods on dense traffic datasets by 30%.

Rohan Chandra, Sachin Grover, Kyungjun Lee et al.

We propose a recurrent variational auto-encoder for texture synthesis. A novel loss function, FLTBNK, is used for training the texture synthesizer. It is rotational and partially color invariant loss function. Unlike L2 loss, FLTBNK explicitly models the correlation of color intensity between pixels. Our texture synthesizer generates neighboring tiles to expand a sample texture and is evaluated using various texture patterns from Describable Textures Dataset (DTD). We perform both quantitative and qualitative experiments with various loss functions to evaluate the performance of our proposed loss function (FLTBNK) --- a mini-human subject study is used for the qualitative evaluation.