Federico Rossi

Ramon Iglesias, Federico Rossi, Rick Zhang et al.

In this paper we present a queueing network approach to the problem of routing and rebalancing a fleet of self-driving vehicles providing on-demand mobility within a capacitated road network. We refer to such systems as autonomous mobility-on-demand systems, or AMoD. We first cast an AMoD system into a closed, multi-class BCMP queueing network model. Second, we present analysis tools that allow the characterization of performance metrics for a given routing policy, in terms, e.g., of vehicle availabilities, and first and second order moments of vehicle throughput. Third, we propose a scalable method for the synthesis of routing policies, with performance guarantees in the limit of large fleet sizes. Finally, we validate our theoretical results on a case study of New York City. Collectively, this paper provides a unifying framework for the analysis and control of AMoD systems, which subsumes earlier Jackson and network flow models, provides a quite large set of modeling options (e.g., the inclusion of road capacities and general travel time distributions), and allows the analysis of second and higher-order moments for the performance metrics.

Federico Rossi, Marco Pavone

In this paper we study the inherent trade-off between time and communication complexity for the distributed consensus problem. In our model, communication complexity is measured as the maximum data throughput (in bits per second) sent through the network at a given instant. Such a notion of communication complexity, referred to as bandwidth complexity, is related to the frequency bandwidth a designer should collectively allocate to the agents if they were to communicate via a wireless channel, which represents an important constraint for dense robotic networks. We prove a lower bound on the bandwidth complexity of the consensus problem and provide a consensus algorithm that is bandwidth-optimal for a wide class of consensus functions. We then propose a distributed algorithm that can trade communication complexity versus time complexity as a function of a tunable parameter, which can be adjusted by a system designer as a function of the properties of the wireless communication channel. We rigorously characterize the tunable algorithm's worst-case bandwidth complexity and show that it compares favorably with the bandwidth complexity of well-known consensus algorithm.

Rick Zhang, Federico Rossi, Marco Pavone

In this paper we present a model predictive control (MPC) approach to optimize vehicle scheduling and routing in an autonomous mobility-on-demand (AMoD) system. In AMoD systems, robotic, self-driving vehicles transport customers within an urban environment and are coordinated to optimize service throughout the entire network. Specifically, we first propose a novel discrete-time model of an AMoD system and we show that this formulation allows the easy integration of a number of real-world constraints, e.g., electric vehicle charging constraints. Second, leveraging our model, we design a model predictive control algorithm for the optimal coordination of an AMoD system and prove its stability in the sense of Lyapunov. At each optimization step, the vehicle scheduling and routing problem is solved as a mixed integer linear program (MILP) where the decision variables are binary variables representing whether a vehicle will 1) wait at a station, 2) service a customer, or 3) rebalance to another station. Finally, by using real-world data, we show that the MPC algorithm can be run in real-time for moderately-sized systems and outperforms previous control strategies for AMoD systems.

Qi Heng Ho, Tyler Becker, Benjamin Kraske et al.

Many sequential decision problems involve optimizing one objective function while imposing constraints on other objectives. Constrained Partially Observable Markov Decision Processes (C-POMDP) model this case with transition uncertainty and partial observability. In this work, we first show that C-POMDPs violate the optimal substructure property over successive decision steps and thus may exhibit behaviors that are undesirable for some (e.g., safety critical) applications. Additionally, online re-planning in C-POMDPs is often ineffective due to the inconsistency resulting from this violation. To address these drawbacks, we introduce the Recursively-Constrained POMDP (RC-POMDP), which imposes additional history-dependent cost constraints on the C-POMDP. We show that, unlike C-POMDPs, RC-POMDPs always have deterministic optimal policies and that optimal policies obey Bellman's principle of optimality. We also present a point-based dynamic programming algorithm for RC-POMDPs. Evaluations on benchmark problems demonstrate the efficacy of our algorithm and show that policies for RC-POMDPs produce more desirable behaviors than policies for C-POMDPs.

Benjamin Kraske, Qi Heng Ho, Federico Rossi et al.

Policies for Partially Observable Markov Decision Processes (POMDPs) are often designed using a nominal system model. In practice, this model can deviate from the true system during deployment due to factors such as calibration drift or sensor degradation, leading to unexpected performance degradation. This work studies policy robustness against deviations in the POMDP observation model. We introduce the Policy Observation Robustness Problem: to determine the maximum tolerable deviation in a POMDP's observation model that guarantees the policy's value remains above a specified threshold. We analyze two variants: the sticky variant, where deviations are dependent on state and actions, and the non-sticky variant, where they can be history-dependent. We show that the Policy Observation Robustness Problem can be formulated as a bi-level optimization problem in which the inner optimization is monotonic in the size of the observation deviation. This enables efficient solutions using root-finding algorithms in the outer optimization. For the non-sticky variant, we show that when policies are represented with finite-state controllers (FSCs) it is sufficient to consider observations which depend on nodes in the FSC rather than full histories. We present Robust Interval Search, an algorithm with soundness and convergence guarantees, for both the sticky and non-sticky variants. We show this algorithm has polynomial time complexity in the non-sticky variant and at most exponential time complexity in the sticky variant. We provide experimental results validating and demonstrating the scalability of implementations of Robust Interval Search to POMDP problems with tens of thousands of states. We also provide case studies from robotics and operations research which demonstrate the practical utility of the problem and algorithms.

Federico Rossi, Marco Cococcioni, Roger Ferrer Ibàñez et al.

As recently demonstrated, Deep Neural Networks (DNN), usually trained using single precision IEEE 754 floating point numbers (binary32), can also work using lower precision. Therefore, 16-bit and 8-bit compressed format have attracted considerable attention. In this paper, we focused on two families of formats that have already achieved interesting results in compressing binary32 numbers in machine learning applications, without sensible degradation of the accuracy: bfloat and posit. Even if 16-bit and 8-bit bfloat/posit are routinely used for reducing the storage of the weights/biases of trained DNNs, the inference still often happens on the 32-bit FPU of the CPU (especially if GPUs are not available). In this paper we propose a way to decompress a tensor of bfloat/posits just before computations, i.e., after the compressed operands have been loaded within the vector registers of a vector capable CPU, in order to save bandwidth usage and increase cache efficiency. Finally, we show the architectural parameters and considerations under which this solution is advantageous with respect to the uncompressed one.

Marco Cococcioni, Dario Pagani, Federico Rossi

The increasing computational and memory demands of large language models (LLMs) necessitate innovative approaches to optimize resource usage without compromising performance. This paper leverages microscaling floating-point formats, a novel technique designed to address these challenges by reducing the storage and computational overhead associated with numerical representations in LLMs. Unlike traditional floating-point representations that allocate a dedicated scale for each value, microscaling employs a shared scale across a block of values, enabling compact one-byte floating-point representations while maintaining an extended dynamic range. We explore the application of microscaling in the context of 8-bit floating-point formats to significantly reduce memory footprint and computational costs. We tested several configurations of microscaling floats within the GPT-2 LLM architecture, demonstrating that microscaling data formats can achieve competitive accuracy during training and inference, proving its efficacy as a resource-efficient alternative for deploying LLMs at scale. The source code is publicly available at: https://github.com/unipi-dii-compressedarith/llm.c-sve

Federico Rossi, Tiago Stegun Vaquero, Marc Sanchez Net et al.

We present the Pluggable Distributed Resource Allocator (PDRA), a middleware for distributed computing in heterogeneous mobile robotic networks. PDRA enables autonomous robotic agents to share computational resources for computationally expensive tasks such as localization and path planning. It sits between an existing single-agent planner/executor and existing computational resources (e.g. ROS packages), intercepts the executor's requests and, if needed, transparently routes them to other robots for execution. PDRA is pluggable: it can be integrated in an existing single-robot autonomy stack with minimal modifications. Task allocation decisions are performed by a mixed-integer programming algorithm, solved in a shared-world fashion, that models CPU resources, latency requirements, and multi-hop, periodic, bandwidth-limited network communications; the algorithm can minimize overall energy usage or maximize the reward for completing optional tasks. Simulation results show that PDRA can reduce energy and CPU usage by over 50% in representative multi-robot scenarios compared to a naive scheduler; runs on embedded platforms; and performs well in delay- and disruption-tolerant networks (DTNs). PDRA is available to the community under an open-source license.

Federico Lozano-Cuadra, Beatriz Soret, Marc Sanchez Net et al.

We present a fully decentralized routing framework for multi-robot exploration missions operating under the constraints of a Lunar Delay-Tolerant Network (LDTN). In this setting, autonomous rovers must relay collected data to a lander under intermittent connectivity and unknown mobility patterns. We formulate the problem as a Partially Observable Markov Decision Problem (POMDP) and propose a Graph Attention-based Multi-Agent Reinforcement Learning (GAT-MARL) policy that performs Centralized Training, Decentralized Execution (CTDE). Our method relies only on local observations and does not require global topology updates or packet replication, unlike classical approaches such as shortest path and controlled flooding-based algorithms. Through Monte Carlo simulations in randomized exploration environments, GAT-MARL provides higher delivery rates, no duplications, and fewer packet losses, and is able to leverage short-term mobility forecasts; offering a scalable solution for future space robotic systems for planetary exploration, as demonstrated by successful generalization to larger rover teams.

Qi Heng Ho, Martin S. Feather, Federico Rossi et al.

Partially Observable Markov Decision Processes (POMDPs) are powerful models for sequential decision making under transition and observation uncertainties. This paper studies the challenging yet important problem in POMDPs known as the (indefinite-horizon) Maximal Reachability Probability Problem (MRPP), where the goal is to maximize the probability of reaching some target states. This is also a core problem in model checking with logical specifications and is naturally undiscounted (discount factor is one). Inspired by the success of point-based methods developed for discounted problems, we study their extensions to MRPP. Specifically, we focus on trial-based heuristic search value iteration techniques and present a novel algorithm that leverages the strengths of these techniques for efficient exploration of the belief space (informed search via value bounds) while addressing their drawbacks in handling loops for indefinite-horizon problems. The algorithm produces policies with two-sided bounds on optimal reachability probabilities. We prove convergence to an optimal policy from below under certain conditions. Experimental evaluations on a suite of benchmarks show that our algorithm outperforms existing methods in almost all cases in both probability guarantees and computation time.

Joshua Vander Hook, Federico Rossi, Tiago Vaquero et al.

Computation load-sharing across a network of heterogeneous robots is a promising approach to increase robots capabilities and efficiency as a team in extreme environments. However, in such environments, communication links may be intermittent and connections to the cloud or internet may be nonexistent. In this paper we introduce a communication-aware, computation task scheduling problem for multi-robot systems and propose an integer linear program (ILP) that optimizes the allocation of computational tasks across a network of heterogeneous robots, accounting for the networked robots' computational capabilities and for available (and possibly time-varying) communication links. We consider scheduling of a set of inter-dependent required and optional tasks modeled by a dependency graph. We present a consensus-backed scheduling architecture for shared-world, distributed systems. We validate the ILP formulation and the distributed implementation in different computation platforms and in simulated scenarios with a bias towards lunar or planetary exploration scenarios. Our results show that the proposed implementation can optimize schedules to allow a threefold increase the amount of rewarding tasks performed (e.g., science measurements) compared to an analogous system with no computational load-sharing.

Rebecca Castano, Tiago Vaquero, Federico Rossi et al.

Onboard autonomy technologies such as planning and scheduling, identification of scientific targets, and content-based data summarization, will lead to exciting new space science missions. However, the challenge of operating missions with such onboard autonomous capabilities has not been studied to a level of detail sufficient for consideration in mission concepts. These autonomy capabilities will require changes to current operations processes, practices, and tools. We have developed a case study to assess the changes needed to enable operators and scientists to operate an autonomous spacecraft by facilitating a common model between the ground personnel and the onboard algorithms. We assess the new operations tools and workflows necessary to enable operators and scientists to convey their desired intent to the spacecraft, and to be able to reconstruct and explain the decisions made onboard and the state of the spacecraft. Mock-ups of these tools were used in a user study to understand the effectiveness of the processes and tools in enabling a shared framework of understanding, and in the ability of the operators and scientists to effectively achieve mission science objectives.

Federico Rossi, Saptarshi Bandyopadhyay, Michael T. Wolf et al.

The goal of this paper is to provide a survey and application-focused atlas of collective behavior coordination algorithms for multi-agent systems. We survey the general family of collective behavior algorithms for multi-agent systems and classify them according to their underlying mathematical structure. In doing so, we aim to capture fundamental mathematical properties of algorithms (e.g., scalability with respect to the number of agents and bandwidth use) and to show how the same algorithm or family of algorithms can be used for multiple tasks and applications. Collectively, this paper provides an application-focused atlas of algorithms for collective behavior of multi-agent systems, with three objectives: 1. to act as a tutorial guide to practitioners in the selection of coordination algorithms for a given application; 2. to highlight how mathematically similar algorithms can be used for a variety of tasks, ranging from low-level control to high-level coordination; 3. to explore the state-of-the-art in the field of control of multi-agent systems and identify areas for future research.

Kiril Solovey, Saptarshi Bandyopadhyay, Federico Rossi et al.

Multi-robot systems are uniquely well-suited to performing complex tasks such as patrolling and tracking, information gathering, and pick-up and delivery problems, offering significantly higher performance than single-robot systems. A fundamental building block in most multi-robot systems is task allocation: assigning robots to tasks (e.g., patrolling an area, or servicing a transportation request) as they appear based on the robots' states to maximize reward. In many practical situations, the allocation must account for heterogeneous capabilities (e.g., availability of appropriate sensors or actuators) to ensure the feasibility of execution, and to promote a higher reward, over a long time horizon. To this end, we present the FlowDec algorithm for efficient heterogeneous task-allocation achieving an approximation factor of at least 1/2 of the optimal reward. Our approach decomposes the heterogeneous problem into several homogeneous subproblems that can be solved efficiently using min-cost flow. Through simulation experiments, we show that our algorithm is faster by several orders of magnitude than a MILP approach.

Suyun Bae, Federico Rossi, Joshua Vander Hook et al.

Autonomous multi-robot systems, where a team of robots shares information to perform tasks that are beyond an individual robot's abilities, hold great promise for a number of applications, such as planetary exploration missions. Each robot in a multi-robot system that uses the shared-world coordination paradigm autonomously schedules which robot should perform a given task, and when, using its worldview--the robot's internal representation of its belief about both its own state, and other robots' states. A key problem for operators is that robots' worldviews can fall out of sync (often due to weak communication links), leading to desynchronization of the robots' scheduling decisions and inconsistent emergent behavior (e.g., tasks not performed, or performed by multiple robots). Operators face the time-consuming and difficult task of making sense of the robots' scheduling decisions, detecting de-synchronizations, and pinpointing the cause by comparing every robot's worldview. To address these challenges, we introduce MOSAIC Viewer, a visual analytics system that helps operators (i) make sense of the robots' schedules and (ii) detect and conduct a root cause analysis of the robots' desynchronized worldviews. Over a year-long partnership with roboticists at the NASA Jet Propulsion Laboratory, we conduct a formative study to identify the necessary system design requirements and a qualitative evaluation with 12 roboticists. We find that MOSAIC Viewer is faster- and easier-to-use than the users' current approaches, and it allows them to stitch low-level details to formulate a high-level understanding of the robots' schedules and detect and pinpoint the cause of the desynchronized worldviews.

Robin Brown, Federico Rossi, Kiril Solovey et al.

A number of prototypical optimization problems in multi-agent systems (e.g., task allocation and network load-sharing) exhibit a highly local structure: that is, each agent's decision variables are only directly coupled to few other agent's variables through the objective function or the constraints. Nevertheless, existing algorithms for distributed optimization generally do not exploit the locality structure of the problem, requiring all agents to compute or exchange the full set of decision variables. In this paper, we develop a rigorous notion of "locality" that quantifies the degree to which agents can compute their portion of the global solution based solely on information in their local neighborhood. This notion provides a theoretical basis for a rather simple algorithm in which agents individually solve a truncated sub-problem of the global problem, where the size of the sub-problem used depends on the locality of the problem, and the desired accuracy. Numerical results show that the proposed theoretical bounds are remarkably tight for well-conditioned problems.

Alvaro Estandia, Maximilian Schiffer, Federico Rossi et al.

In future transportation systems, the charging behavior of electric Autonomous Mobility on Demand (AMoD) fleets, i.e., fleets of electric self-driving cars that service on-demand trip requests, will likely challenge power distribution networks (PDNs), causing overloads or voltage drops. In this paper, we show that these challenges can be significantly attenuated if the PDNs' operational constraints and exogenous loads (e.g., from homes or businesses) are accounted for when operating an electric AMoD fleet. We focus on a system-level perspective, assuming full coordination between the AMoD and the PDN operators. From this single entity perspective, we assess potential coordination benefits. Specifically, we extend previous results on an optimization-based modeling approach for electric AMoD systems to jointly control an electric AMoD fleet and a series of PDNs, and analyze the benefit of coordination under load balancing constraints. For a case study of Orange County, CA, we show that the coordination between the electric AMoD fleet and the PDNs eliminates 99% of the overloads and 50% of the voltage drops that the electric AMoD fleet would cause in an uncoordinated setting. Our results show that coordinating electric AMoD and PDNs can help maintain the reliability of PDNs under added electric AMoD charging load, thus significantly mitigating or deferring the need for PDN capacity upgrades.

Mauro Salazar, Federico Rossi, Maximilian Schiffer et al.

In this paper we study models and coordination policies for intermodal Autonomous Mobility-on-Demand (AMoD), wherein a fleet of self-driving vehicles provides on-demand mobility jointly with public transit. Specifically, we first present a network flow model for intermodal AMoD, where we capture the coupling between AMoD and public transit and the goal is to maximize social welfare. Second, leveraging such a model, we design a pricing and tolling scheme that allows to achieve the social optimum under the assumption of a perfect market with selfish agents. Finally, we present a real-world case study for New York City. Our results show that the coordination between AMoD fleets and public transit can yield significant benefits compared to an AMoD system operating in isolation.

Federico Rossi, Saptarshi Bandyopadhyay, Michael Wolf et al.

In this paper, we review multi-agent collective behavior algorithms in the literature and classify them according to their underlying mathematical structure. For each mathematical technique, we identify the multi-agent coordination tasks it can be applied to, and we analyze its scalability, bandwidth use, and demonstrated maturity. We highlight how versatile techniques such as artificial potential functions can be used for applications ranging from low-level position control to high-level coordination and task allocation, we discuss possible reasons for the slow adoption of complex distributed coordination algorithms in the field, and we highlight areas for further research and development.

Ramon Iglesias, Federico Rossi, Kevin Wang et al.

The goal of this paper is to present an end-to-end, data-driven framework to control Autonomous Mobility-on-Demand systems (AMoD, i.e. fleets of self-driving vehicles). We first model the AMoD system using a time-expanded network, and present a formulation that computes the optimal rebalancing strategy (i.e., preemptive repositioning) and the minimum feasible fleet size for a given travel demand. Then, we adapt this formulation to devise a Model Predictive Control (MPC) algorithm that leverages short-term demand forecasts based on historical data to compute rebalancing strategies. We test the end-to-end performance of this controller with a state-of-the-art LSTM neural network to predict customer demand and real customer data from DiDi Chuxing: we show that this approach scales very well for large systems (indeed, the computational complexity of the MPC algorithm does not depend on the number of customers and of vehicles in the system) and outperforms state-of-the-art rebalancing strategies by reducing the mean customer wait time by up to to 89.6%.

Federico Rossi, Ramon Iglesias, Mahnoosh Alizadeh et al.

We study the interaction between a fleet of electric, self-driving vehicles servicing on-demand transportation requests (referred to as Autonomous Mobility-on-Demand, or AMoD, system) and the electric power network. We propose a model that captures the coupling between the two systems stemming from the vehicles' charging requirements and captures time-varying customer demand and power generation costs, road congestion, battery depreciation, and power transmission and distribution constraints. We then leverage the model to jointly optimize the operation of both systems. We devise an algorithmic procedure to losslessly reduce the problem size by bundling customer requests, allowing it to be efficiently solved by off-the-shelf linear programming solvers. Next, we show that the socially optimal solution to the joint problem can be enforced as a general equilibrium, and we provide a dual decomposition algorithm that allows self-interested agents to compute the market clearing prices without sharing private information. We assess the performance of the mode by studying a hypothetical AMoD system in Dallas-Fort Worth and its impact on the Texas power network. Lack of coordination between the AMoD system and the power network can cause a 4.4% increase in the price of electricity in Dallas-Fort Worth; conversely, coordination between the AMoD system and the power network could reduce electricity expenditure compared to the case where no cars are present (despite the increased demand for electricity) and yield savings of up $147M/year. Finally, we provide a receding-horizon implementation and assess its performance with agent-based simulations. Collectively, the results of this paper provide a first-of-a-kind characterization of the interaction between electric-powered AMoD systems and the power network, and shed additional light on the economic and societal value of AMoD.

Federico Rossi, Rick Zhang, Marco Pavone

In this paper we study the routing and rebalancing problem for a fleet of autonomous vehicles providing on-demand transportation within a congested urban road network (that is, a road network where traffic speed depends on vehicle density). We show that the congestion-free routing and rebalancing problem is NP-hard and provide a randomized algorithm which finds a low-congestion solution to the routing and rebalancing problem that approximately minimizes the number of vehicles on the road in polynomial time. We provide theoretical bounds on the probability of violating the congestion constraints; we also characterize the expected number of vehicles required by the solution with a commonly-used empirical congestion model and provide a bound on the approximation factor of the algorithm. Numerical experiments on a realistic road network with real-world customer demands show that our algorithm introduces very small amounts of congestion. The performance of our algorithm in terms of travel times and required number of vehicles is very close to (and sometimes better than) the optimal congestion-free solution.

Federico Rossi, Marco Pavone

This paper studies fundamental limitations of performance for distributed decision-making in robotic networks. The class of decision-making problems we consider encompasses a number of prototypical problems such as average-based consensus as well as distributed optimization, leader election, majority voting, MAX, MIN, and logical formulas. We first propose a formal model for distributed computation on robotic networks that is based on the concept of I/O automata and is inspired by the Computer Science literature on distributed computing clusters. Then, we present a number of bounds on time, message, and byte complexity, which we use to discuss the relative performance of a number of approaches for distributed decision-making. From a methodological standpoint, our work sheds light on the relation between the tools developed by the Computer Science and Controls communities on the topic of distributed algorithms.