Christoffer Heckman

Christoffer Heckman, Harel Biggie, Brendan Crowe et al.

Persistent maps used by autonomous robots increasingly fuse a geometric perception stack whose assertions are well-characterized with a foundation-model channel that produces semantic claims without calibrated reliability about the same scene. Contemporary mapping systems integrate the two channels by treating the foundation-model channel as an additional voter into a per-element posterior, uncalibrated for its own per-class reliability and without machinery to flag when the two channels contradict each other at a given moment. We propose an update operator with two cooperating mechanisms: a per-class calibrated commit gate, and a per-event conflict-drop window that refuses to commit foundation-model claims contradicted by the geometric channel at the moment of the claim. We evaluate on KITTI-360 and ScanNet, with an oracle geometric channel (panoptic ground truth) and an off-the-shelf online semantic segmenter (Mask2Former) to demonstrate real-world performance. The operator produces substantially more accurate committed maps (KITTI is car commit precision 99.7% vs. 43.9% for the calibration-only operator; mean per-class IoU 0.522 vs. 0.180), retains more compositional true positives at higher precision than a monolithic compositional VLM prompt. The framework operates at deployment quality across both oracle and off-the-shelf-segmenter geometric channels, and is invariant under foundation-model substitution.

Harel Biggie, Andrew Beathard, Christoffer Heckman

Typical algorithms for point cloud registration such as Iterative Closest Point (ICP) require a favorable initial transform estimate between two point clouds in order to perform a successful registration. State-of-the-art methods for choosing this starting condition rely on stochastic sampling or global optimization techniques such as branch and bound. In this work, we present a new method based on Bayesian optimization for finding the critical initial ICP transform. We provide three different configurations for our method which highlights the versatility of the algorithm to both find rapid results and refine them in situations where more runtime is available such as offline map building. Experiments are run on popular data sets and we show that our approach outperforms state-of-the-art methods when given similar computation time. Furthermore, it is compatible with other improvements to ICP, as it focuses solely on the selection of an initial transform, a starting point for all ICP-based methods.

Eugene R. Rush, Christoffer Heckman, Kaushik Jayaram et al.

Recurrent neural network-based reinforcement learning systems are capable of complex motor control tasks such as locomotion and manipulation, however, much of their underlying mechanisms still remain difficult to interpret. Our aim is to leverage computational neuroscience methodologies to understanding the population-level activity of robust robot locomotion controllers. Our investigation begins by analyzing topological structure, discovering that fragile controllers have a higher number of fixed points with unstable directions, resulting in poorer balance when instructed to stand in place. Next, we analyze the forced response of the system by applying targeted neural perturbations along directions of dominant population-level activity. We find evidence that recurrent state dynamics are structured and low-dimensional during walking, which aligns with primate studies. Additionally, when recurrent states are perturbed to zero, fragile agents continue to walk, which is indicative of a stronger reliance on sensory input and weaker recurrence.

Xuefei Sun, Xujia Zhang, Brendan Crowe et al.

Zero-shot 3D visual grounding requires localizing objects in unstructured environments from free-form natural language. Recent vision-language model (VLM) approaches achieve promising results but rely on view-dependent reasoning or implicit representations, limiting spatial consistency and interpretability for compositional queries. We propose SceneGraphGrounder, a framework that reformulates 3D grounding as structured graph matching over a reconstructed 3D scene graph. To enable this formulation, we introduce a visual marker prompting strategy that enables a VLM to infer object-object relationships from 2D views, which are subsequently lifted into a persistent 3D scene graph encoding both spatial and semantic relations. Given a query, we construct a query graph and perform constrained alignment with the scene graph, ensuring multi-view consistency and interpretable reasoning. Experiments on the ScanRefer benchmark demonstrate that our method achieves competitive performance among zero-shot approaches, using only RGB-D inputs. We further validate our framework through real-world deployment on a mobile robot, demonstrating robust spatial reasoning in long-horizon physical environments. We will make our code publicly available upon acceptance.

Kali Hamilton, Christoffer Heckman

Cameras and LiDAR degrade in rain, fog, and snow, while millimeter-wave radar remains largely unaffected. We align a radar encoder to frozen SigLIP vision embeddings and decode structured scene captions through a frozen vision-language model (VLM) with approximately 7M trainable parameters. On K-RADAR with held-out fog, light snow, and heavy snow sequences, all radar configurations outperform a camera baseline that collapses to over 90% hallucination. We identify a token-norm mismatch as the dominant failure mode when bridging radar to a frozen VLM and show that projector-output LayerNorm resolves it. Analysis of encoder complexity, caption format, and pooling strategy reveals tradeoffs that inform future radar-VLM pipeline design.

Lorin Achey, Alec Reed, Brendan Crowe et al.

We present a novel approach for enhancing robotic exploration by using generative occupancy mapping. We introduce SceneSense, a diffusion model designed and trained for predicting 3D occupancy maps given partial observations. Our proposed approach probabilistically fuses these predictions into a running occupancy map in real-time, resulting in significant improvements in map quality and traversability. We implement SceneSense onboard a quadruped robot and validate its performance with real-world experiments to demonstrate the effectiveness of the model. In these experiments, we show that occupancy maps enhanced with SceneSense predictions better represent our fully observed ground truth data (24.44% FID improvement around the robot and 75.59% improvement at range). We additionally show that integrating SceneSense-enhanced maps into our robotic exploration stack as a "drop-in" map improvement, utilizing an existing off-the-shelf planner, results in improvements in robustness and traversability time. Finally we show results of full exploration evaluations with our proposed system in two dissimilar environments and find that locally enhanced maps provide more consistent exploration results than maps constructed only from direct sensor measurements.

Xujia Zhang, Brendan Crowe, Christoffer Heckman

The completion, extension, and generation of 3D semantic scenes are an interrelated set of capabilities that are useful for robotic navigation and exploration. Existing approaches seek to decouple these problems and solve them oneoff. Additionally, these approaches are often domain-specific, requiring separate models for different data distributions, e.g. indoor vs. outdoor scenes. To unify these techniques and provide cross-domain compatibility, we develop a single framework that can perform scene completion, extension, and generation in both indoor and outdoor scenes, which we term Octree Latent Semantic Diffusion. Our approach operates directly on an efficient dual octree graph latent representation: a hierarchical, sparse, and memory-efficient occupancy structure. This technique disentangles synthesis into two stages: (i) structure diffusion, which predicts binary split signals to construct a coarse occupancy octree, and (ii) latent semantic diffusion, which generates semantic embeddings decoded by a graph VAE into voxellevel semantic labels. To perform semantic scene completion or extension, our model leverages inference-time latent inpainting, or outpainting respectively. These inference-time methods use partial LiDAR scans or maps to condition generation, without the need for retraining or finetuning. We demonstrate highquality structure, coherent semantics, and robust completion from single LiDAR scans, as well as zero-shot generalization to out-of-distribution LiDAR data. These results indicate that completion-through-generation in a dual octree graph latent space is a practical and scalable alternative to regression-based pipelines for real-world robotic perception tasks.

Alec Farid, Peter Schleede, Aaron Huang et al.

We are motivated by the problem of autonomous vehicle performance validation. A key challenge is that an autonomous vehicle requires testing in every kind of driving scenario it could encounter, including rare events, to provide a strong case for safety and show there is no edge-case pathological behavior. Autonomous vehicle companies rely on potentially millions of miles driven in realistic simulation to expose the driving stack to enough miles to estimate rates and severity of collisions. To address scalability and coverage, we propose the use of a behavior foundation model, specifically a masked autoencoder (MAE), trained to reconstruct driving scenarios. We leverage the foundation model in two complementary ways: we (i) use the learned embedding space to group qualitatively similar scenarios together and (ii) fine-tune the model to label scenario difficulty based on the likelihood of a collision upon simulation. We use the difficulty scoring as importance weighting for the groups of scenarios. The result is an approach which can more rapidly estimate the rates and severity of collisions by prioritizing hard scenarios while ensuring exposure to every kind of driving scenario.

Michael T. Ohradzansky, Eugene R. Rush, Danny G. Riley et al.

Artificial intelligence has undergone immense growth and maturation in recent years, though autonomous systems have traditionally struggled when fielded in diverse and previously unknown environments. DARPA is seeking to change that with the Subterranean Challenge, by providing roboticists the opportunity to support civilian and military first responders in complex and high-risk underground scenarios. The subterranean domain presents a handful of challenges, such as limited communication, diverse topology and terrain, and degraded sensing. Team MARBLE proposes a solution for autonomous exploration of unknown subterranean environments in which coordinated agents search for artifacts of interest. The team presents two navigation algorithms in the form of a metric-topological graph-based planner and a continuous frontier-based planner. To facilitate multi-agent coordination, agents share and merge new map information and candidate goal-points. Agents deploy communication beacons at different points in the environment, extending the range at which maps and other information can be shared. Onboard autonomy reduces the load on human supervisors, allowing agents to detect and localize artifacts and explore autonomously outside established communication networks. Given the scale, complexity, and tempo of this challenge, a range of lessons were learned, most importantly, that frequent and comprehensive field testing in representative environments is key to rapidly refining system performance.

Andrew Kramer, Kyle Harlow, Christopher Williams et al.

Millimeter wave radar is becoming increasingly popular as a sensing modality for robotic mapping and state estimation. However, there are very few publicly available datasets that include dense, high-resolution millimeter wave radar scans and there are none focused on 3D odometry and mapping. In this paper we present a solution to that problem. The ColoRadar dataset includes 3 different forms of dense, high-resolution radar data from 2 FMCW radar sensors as well as 3D lidar, IMU, and highly accurate groundtruth for the sensor rig's pose over approximately 2 hours of data collection in highly diverse 3D environments.

Mike Kasper, Fernando Nobre, Christoffer Heckman et al.

Training networks to perform metric relocalization traditionally requires accurate image correspondences. In practice, these are obtained by restricting domain coverage, employing additional sensors, or capturing large multi-view datasets. We instead propose a self-supervised solution, which exploits a key insight: localizing a query image within a map should yield the same absolute pose, regardless of the reference image used for registration. Guided by this intuition, we derive a novel transform consistency loss. Using this loss function, we train a deep neural network to infer dense feature and saliency maps to perform robust metric relocalization in dynamic environments. We evaluate our framework on synthetic and real-world data, showing our approach outperforms other supervised methods when a limited amount of ground-truth information is available.

Joewie J. Koh, Guohui Ding, Christoffer Heckman et al.

Multi-robot cooperation requires agents to make decisions that are consistent with the shared goal without disregarding action-specific preferences that might arise from asymmetry in capabilities and individual objectives. To accomplish this goal, we propose a method named SLiCC: Stackelberg Learning in Cooperative Control. SLiCC models the problem as a partially observable stochastic game composed of Stackelberg bimatrix games, and uses deep reinforcement learning to obtain the payoff matrices associated with these games. Appropriate cooperative actions are then selected with the derived Stackelberg equilibria. Using a bi-robot cooperative object transportation problem, we validate the performance of SLiCC against centralized multi-agent Q-learning and demonstrate that SLiCC achieves better combined utility.

Samuel Bateman, Kyle Harlow, Christoffer Heckman

Many modern simultaneous localization and mapping (SLAM) techniques rely on sparse landmark-based maps due to their real-time performance. However, these techniques frequently assert that these landmarks are fixed in position over time, known as the static-world assumption. This is rarely, if ever, the case in most real-world environments. Even worse, over long deployments, robots are bound to observe traditionally static landmarks change, for example when an autonomous vehicle encounters a construction zone. This work addresses this challenge, accounting for changes in complex three-dimensional environments with the creation of a probabilistic filter that operates on the features that give rise to landmarks. To accomplish this, landmarks are clustered into cliques and a filter is developed to estimate their persistence jointly among observations of the landmarks in a clique. This filter uses estimated spatial-temporal priors of geometric objects, allowing for dynamic and semi-static objects to be removed from a formally static map. The proposed algorithm is validated in a 3D simulated environment.

Hadas Kress-Gazit, Kerstin Eder, Guy Hoffman et al.

Robot capabilities are maturing across domains, from self-driving cars, to bipeds and drones. As a result, robots will soon no longer be confined to safety-controlled industrial settings; instead, they will directly interact with the general public. The growing field of Human-Robot Interaction (HRI) studies various aspects of this scenario - from social norms to joint action to human-robot teams and more. Researchers in HRI have made great strides in developing models, methods, and algorithms for robots acting with and around humans, but these "computational HRI" models and algorithms generally do not come with formal guarantees and constraints on their operation. To enable human-interactive robots to move from the lab to real-world deployments, we must address this gap. This article provides an overview of verification, validation and synthesis techniques used to create demonstrably trustworthy systems, describes several HRI domains that could benefit from such techniques, and provides a roadmap for the challenges and the research needed to create formalized and guaranteed human-robot interaction.

Guohui Ding, Joewie J. Koh, Kelly Merckaert et al.

We consider solving a cooperative multi-robot object manipulation task using reinforcement learning (RL). We propose two distributed multi-agent RL approaches: distributed approximate RL (DA-RL), where each agent applies Q-learning with individual reward functions; and game-theoretic RL (GT-RL), where the agents update their Q-values based on the Nash equilibrium of a bimatrix Q-value game. We validate the proposed approaches in the setting of cooperative object manipulation with two simulated robot arms. Although we focus on a small system of two agents in this paper, both DA-RL and GT-RL apply to general multi-agent systems, and are expected to scale well to large systems.

Zhaozhong Chen, Nisar Ahmed, Simon Julier et al.

Many state estimation algorithms must be tuned given the state space process and observation models, the process and observation noise parameters must be chosen. Conventional tuning approaches rely on heuristic hand-tuning or gradient-based optimization techniques to minimize a performance cost function. However, the relationship between tuned noise values and estimator performance is highly nonlinear and stochastic. Therefore, the tuning solutions can easily get trapped in local minima, which can lead to poor choices of noise parameters and suboptimal estimator performance. This paper describes how Bayesian Optimization (BO) can overcome these issues. BO poses optimization as a Bayesian search problem for a stochastic ``black box'' cost function, where the goal is to search the solution space to maximize the probability of improving the current best solution. As such, BO offers a principled approach to optimization-based estimator tuning in the presence of local minima and performance stochasticity. While extended Kalman filters (EKFs) are the main focus of this work, BO can be similarly used to tune other related state space filters. The method presented here uses performance metrics derived from normalized innovation squared (NIS) filter residuals obtained via sensor data, which renders knowledge of ground-truth states unnecessary. The robustness, accuracy, and reliability of BO-based tuning is illustrated on practical nonlinear state estimation problems,losed-loop aero-robotic control.

Fernando Nobre, Christoffer Heckman

We propose a solution for sensor extrinsic self-calibration with very low time complexity, competitive accuracy and graceful handling of often-avoided corner cases: drift in calibration parameters and unobservable directions in the parameter space. It consists of three main parts: 1) information-theoretic based segment selection for constant-time estimation; 2) observability-aware parameter update through a rank-revealing decomposition of the Fisher information matrix; 3) drift-correcting self-calibration through the time-decay of segments. At the core of our FastCal algorithm is the loosely-coupled formulation for sensor extrinsics calibration and efficient selection of measurements. FastCal runs up to an order of magnitude faster than similar self-calibration algorithms (camera-to-camera extrinsics, excluding feature-matching and image pre-processing on all comparisons), making FastCal ideal for integration into existing, resource-constrained, robotics systems.

Steve McGuire, Christoffer Heckman, Daniel Szafir et al.

Determining extrinsic calibration parameters is a necessity in any robotic system composed of actuators and cameras. Once a system is outside the lab environment, parameters must be determined without relying on outside artifacts such as calibration targets. We propose a method that relies on structured motion of an observed arm to recover extrinsic calibration parameters. Our method combines known arm kinematics with observations of conics in the image plane to calculate maximum-likelihood estimates for calibration extrinsics. This method is validated in simulation and tested against a real-world model, yielding results consistent with ruler-based estimates. Our method shows promise for estimating the pose of a camera relative to an articulated arm's end effector without requiring tedious measurements or external artifacts. Index Terms: robotics, hand-eye problem, self-calibration, structure from motion

Zhaozhong Chen, Christoffer Heckman, Simon Julier et al.

Kalman filters are routinely used for many data fusion applications including navigation, tracking, and simultaneous localization and mapping problems. However, significant time and effort is frequently required to tune various Kalman filter model parameters, e.g. process noise covariance, pre-whitening filter models for non-white noise, etc. Conventional optimization techniques for tuning can get stuck in poor local minima and can be expensive to implement with real sensor data. To address these issues, a new "black box" Bayesian optimization strategy is developed for automatically tuning Kalman filters. In this approach, performance is characterized by one of two stochastic objective functions: normalized estimation error squared (NEES) when ground truth state models are available, or the normalized innovation error squared (NIS) when only sensor data is available. By intelligently sampling the parameter space to both learn and exploit a nonparametric Gaussian process surrogate function for the NEES/NIS costs, Bayesian optimization can efficiently identify multiple local minima and provide uncertainty quantification on its results.

Hadi Ravanbakhsh, Sina Aghli, Christoffer Heckman et al.

We present an approach to path following using so-called control funnel functions. Synthesizing controllers to "robustly" follow a reference trajectory is a fundamental problem for autonomous vehicles. Robustness, in this context, requires our controllers to handle a specified amount of deviation from the desired trajectory. Our approach considers a timing law that describes how fast to move along a given reference trajectory and a control feedback law for reducing deviations from the reference. We synthesize both feedback laws using "control funnel functions" that jointly encode the control law as well as its correctness argument over a mathematical model of the vehicle dynamics. We adapt a previously described demonstration-based learning algorithm to synthesize a control funnel function as well as the associated feedback law. We implement this law on top of a 1/8th scale autonomous vehicle called the Parkour car. We compare the performance of our path following approach against a trajectory tracking approach by specifying trajectories of varying lengths and curvatures. Our experiments demonstrate the improved robustness obtained from the use of control funnel functions.

John Stechschulte, Christoffer Heckman

When registering point clouds resolved from an underlying 2-D pixel structure, such as those resulting from structured light and flash LiDAR sensors, or stereo reconstruction, it is expected that some points in one cloud do not have corresponding points in the other cloud, and that these would occur together, such as along an edge of the depth map. In this work, a hidden Markov random field model is used to capture this prior within the framework of the iterative closest point algorithm. The EM algorithm is used to estimate the distribution parameters and the hidden component memberships. Experiments are presented demonstrating that this method outperforms several other outlier rejection methods when the point clouds have low or moderate overlap.

Nikolaus Correll, Christoffer Heckman

We posit that embodied artificial intelligence is not only a computational, but also a materials problem. While the importance of material and structural properties in the control loop are well understood, materials can take an active role during control by tight integration of sensors, actuators, computation and communication. We envision such materials to abstract functionality, therefore making the construction of intelligent robots more straightforward and robust. For example, robots could be made of bones that measure load, muscles that move, skin that provides the robot with information about the kind and location of tactile sensations ranging from pressure, to texture and damage, eyes that extract high-level information, and brain material that provides computation in a scalable manner. Such materials will not resemble any existing engineered materials, but rather the heterogeneous components out of which their natural counterparts are made. We describe the state-of-the-art in so-called "robotic materials", their opportunities for revolutionizing applications ranging from manipulation to autonomous driving, and open challenges the robotics community needs to address in collaboration with allies, such as wireless sensor network researchers and polymer scientists.

Mike Kasper, Nima Keivan, Gabe Sibley et al.

We present a novel algorithm for light source estimation in scenes reconstructed with a RGB-D camera based on an analytically-derived formulation of path-tracing. Our algorithm traces the reconstructed scene with a custom path-tracer and computes the analytical derivatives of the light transport equation from principles in optics. These derivatives are then used to perform gradient descent, minimizing the photometric error between one or more captured reference images and renders of our current lighting estimation using an environment map parameterization for light sources. We show that our approach of modeling all light sources as points at infinity approximates lights located near the scene with surprising accuracy. Due to the analytical formulation of derivatives, optimization to the solution is considerably accelerated. We verify our algorithm using both real and synthetic data.