CROSS: A Mixture-of-Experts Reinforcement Learning Framework for Generalizable Large-Scale Traffic Signal Control

Recent advances in robotics, automation, and artificial intelligence have enabled urban traffic systems to operate with increasing autonomy towards future smart cities, powered in part by the development of adaptive traffic signal control (ATSC), which dynamically optimizes signal phases to mitigate congestion and optimize traffic. However, achieving effective and generalizable large-scale ATSC remains a significant challenge due to the diverse intersection topologies and highly dynamic, complex traffic demand patterns across the network. Existing RL-based methods typically use a single shared policy for all scenarios, whose limited representational capacity makes it difficult to capture diverse traffic dynamics and generalize to unseen environments. To address these challenges, we propose CROSS, a novel Mixture-of-Experts (MoE)-based decentralized RL framework for generalizable ATSC. We first introduce a Predictive Contrastive Clustering (PCC) module that forecasts short-term state transitions to identify latent traffic patterns, followed by clustering and contrastive learning to enhance pattern-level representation. We further design a Scenario-Adaptive MoE module that augments a shared policy with multiple experts, thus enabling adaptive specialization and more flexible scenario-specific strategies. We conduct extensive experiments in the SUMO simulator on both synthetic and real-world traffic datasets. Compared with state-of-the-art baselines, CROSS achieves superior performance and generalization through improved representation of diverse traffic scenarios.

COIN: Collaborative Interaction-Aware Multi-Agent Reinforcement Learning for Self-Driving Systems

Multi-Agent Self-Driving (MASD) systems provide an effective solution for coordinating autonomous vehicles to reduce congestion and enhance both safety and operational efficiency in future intelligent transportation systems. Multi-Agent Reinforcement Learning (MARL) has emerged as a promising approach for developing advanced end-to-end MASD systems. However, achieving efficient and safe collaboration in dynamic MASD systems remains a significant challenge in dense scenarios with complex agent interactions. To address this challenge, we propose a novel collaborative(CO-) interaction-aware(-IN) MARL framework, named COIN. Specifically, we develop a new counterfactual individual-global twin delayed deep deterministic policy gradient (CIG-TD3) algorithm, crafted in a "centralized training, decentralized execution" (CTDE) manner, which aims to jointly optimize the individual objectives (navigation) and the global objectives (collaboration) of agents. We further introduce a dual-level interaction-aware centralized critic architecture that captures both local pairwise interactions and global system-level dependencies, enabling more accurate global value estimation and improved credit assignment for collaborative policy learning. We conduct extensive simulation experiments in dense urban traffic environments, which demonstrate that COIN consistently outperforms other advanced baseline methods in both safety and efficiency across various system sizes. These results highlight its superiority in complex and dynamic MASD scenarios, as further validated through real-world robot demonstrations. Supplementary videos are available at https://marmotlab.github.io/COIN/

CoordLight: Learning Decentralized Coordination for Network-Wide Traffic Signal Control

Adaptive traffic signal control (ATSC) is crucial in alleviating congestion, maximizing throughput and promoting sustainable mobility in ever-expanding cities. Multi-Agent Reinforcement Learning (MARL) has recently shown significant potential in addressing complex traffic dynamics, but the intricacies of partial observability and coordination in decentralized environments still remain key challenges in formulating scalable and efficient control strategies. To address these challenges, we present CoordLight, a MARL-based framework designed to improve intra-neighborhood traffic by enhancing decision-making at individual junctions (agents), as well as coordination with neighboring agents, thereby scaling up to network-level traffic optimization. Specifically, we introduce the Queue Dynamic State Encoding (QDSE), a novel state representation based on vehicle queuing models, which strengthens the agents' capability to analyze, predict, and respond to local traffic dynamics. We further propose an advanced MARL algorithm, named Neighbor-aware Policy Optimization (NAPO). It integrates an attention mechanism that discerns the state and action dependencies among adjacent agents, aiming to facilitate more coordinated decision-making, and to improve policy learning updates through robust advantage calculation. This enables agents to identify and prioritize crucial interactions with influential neighbors, thus enhancing the targeted coordination and collaboration among agents. Through comprehensive evaluations against state-of-the-art traffic signal control methods over three real-world traffic datasets composed of up to 196 intersections, we empirically show that CoordLight consistently exhibits superior performance across diverse traffic networks with varying traffic flows. The code is available at https://github.com/marmotlab/CoordLight

LATS: Large Language Model Assisted Teacher-Student Framework for Multi-Agent Reinforcement Learning in Traffic Signal Control

Adaptive Traffic Signal Control (ATSC) aims to optimize traffic flow and minimize delays by adjusting traffic lights in real time. Recent advances in Multi-agent Reinforcement Learning (MARL) have shown promise for ATSC, yet existing approaches still suffer from limited representational capacity, often leading to suboptimal performance and poor generalization in complex and dynamic traffic environments. On the other hand, Large Language Models (LLMs) excel at semantic representation, reasoning, and analysis, yet their propensity for hallucination and slow inference speeds often hinder their direct application to decision-making tasks. To address these challenges, we propose a novel learning paradigm named LATS that integrates LLMs and MARL, leveraging the former's strong prior knowledge and inductive abilities to enhance the latter's decision-making process. Specifically, we introduce a plug-and-play teacher-student learning module, where a trained embedding LLM serves as a teacher to generate rich semantic features that capture each intersection's topology structures and traffic dynamics. A much simpler (student) neural network then learns to emulate these features through knowledge distillation in the latent space, enabling the final model to operate independently from the LLM for downstream use in the RL decision-making process. This integration significantly enhances the overall model's representational capacity across diverse traffic scenarios, thus leading to more efficient and generalizable control strategies. Extensive experiments across diverse traffic datasets empirically demonstrate that our method enhances the representation learning capability of RL models, thereby leading to improved overall performance and generalization over both traditional RL and LLM-only approaches. [...]

CAMO: A Conditional Neural Solver for the Multi-objective Multiple Traveling Salesman Problem

Robotic systems often require a team of robots to collectively visit multiple targets while optimizing competing objectives, such as total travel cost and makespan. This setting can be formulated as the Multi-Objective Multiple Traveling Salesman Problem (MOMTSP). Although learning-based methods have shown strong performance on the single-agent TSP and multi-objective TSP variants, they rarely address the combined challenges of multi-agent coordination and multi-objective trade-offs, which introduce dual sources of complexity. To bridge this gap, we propose CAMO, a conditional neural solver for MOMTSP that generalizes across varying numbers of targets, agents, and preference vectors, and yields high-quality approximations to the Pareto front (PF). Specifically, CAMO consists of a conditional encoder to fuse preferences into instance representations, enabling explicit control over multi-objective trade-offs, and a collaborative decoder that coordinates all agents by alternating agent selection and node selection to construct multi-agent tours autoregressively. To further improve generalization, we train CAMO with a REINFORCE-based objective over a mixed distribution of problem sizes. Extensive experiments show that CAMO outperforms both neural and conventional heuristics, achieving a closer approximation of PFs. In addition, ablation results validate the contributions of CAMO's key components, and real-world tests on a mobile robot platform demonstrate its practical applicability.

ImagiNav: Scalable Embodied Navigation via Generative Visual Prediction and Inverse Dynamics

Enabling robots to navigate open-world environments via natural language is critical for general-purpose autonomy. Yet, Vision-Language Navigation has relied on end-to-end policies trained on expensive, embodiment-specific robot data. While recent foundation models trained on vast simulation data show promise, the challenge of scaling and generalizing due to the limited scene diversity and visual fidelity in simulation persists. To address this gap, we propose ImagiNav, a novel modular paradigm that decouples visual planning from robot actuation, enabling the direct utilization of diverse in-the-wild navigation videos. Our framework operates as a hierarchy: a Vision-Language Model first decomposes instructions into textual subgoals; a finetuned generative video model then imagines the future video trajectory towards that subgoal; finally, an inverse dynamics model extracts the trajectory from the imagined video, which can then be tracked via a low-level controller. We additionally develop a scalable data pipeline of in-the-wild navigation videos auto-labeled via inverse dynamics and a pretrained Vision-Language Model. ImagiNav demonstrates strong zero-shot transfer to robot navigation without requiring robot demonstrations, paving the way for generalist robots that learn navigation directly from unlabeled, open-world data.

To Move or Not to Move: Constraint-based Planning Enables Zero-Shot Generalization for Interactive Navigation

Visual navigation typically assumes the existence of at least one obstacle-free path between start and goal, which must be discovered/planned by the robot. However, in real-world scenarios, such as home environments and warehouses, clutter can block all routes. Targeted at such cases, we introduce the Lifelong Interactive Navigation problem, where a mobile robot with manipulation abilities can move clutter to forge its own path to complete sequential object- placement tasks - each involving placing an given object (eg. Alarm clock, Pillow) onto a target object (eg. Dining table, Desk, Bed). To address this lifelong setting - where effects of environment changes accumulate and have long-term effects - we propose an LLM-driven, constraint-based planning framework with active perception. Our framework allows the LLM to reason over a structured scene graph of discovered objects and obstacles, deciding which object to move, where to place it, and where to look next to discover task-relevant information. This coupling of reasoning and active perception allows the agent to explore the regions expected to contribute to task completion rather than exhaustively mapping the environment. A standard motion planner then executes the corresponding navigate-pick-place, or detour sequence, ensuring reliable low-level control. Evaluated in physics-enabled ProcTHOR-10k simulator, our approach outperforms non-learning and learning-based baselines. We further demonstrate our approach qualitatively on real-world hardware.

GPO: Growing Policy Optimization for Legged Robot Locomotion and Whole-Body Control

Training reinforcement learning (RL) policies for legged robots remains challenging due to high-dimensional continuous actions, hardware constraints, and limited exploration. Existing methods for locomotion and whole-body control work well for position-based control with environment-specific heuristics (e.g., reward shaping, curriculum design, and manual initialization), but are less effective for torque-based control, where sufficiently exploring the action space and obtaining informative gradient signals for training is significantly more difficult. We introduce Growing Policy Optimization (GPO), a training framework that applies a time-varying action transformation to restrict the effective action space in the early stage, thereby encouraging more effective data collection and policy learning, and then progressively expands it to enhance exploration and achieve higher expected return. We prove that this transformation preserves the PPO update rule and introduces only bounded, vanishing gradient distortion, thereby ensuring stable training. We evaluate GPO on both quadruped and hexapod robots, including zero-shot deployment of simulation-trained policies on hardware. Policies trained with GPO consistently achieve better performance. These results suggest that GPO provides a general, environment-agnostic optimization framework for learning legged locomotion.

FARE: Fast-Slow Agentic Robotic Exploration

This work advances autonomous robot exploration by integrating agent-level semantic reasoning with fast local control. We introduce FARE, a hierarchical autonomous exploration framework that integrates a large language model (LLM) for global reasoning with a reinforcement learning (RL) policy for local decision making. FARE follows a fast-slow thinking paradigm. The slow-thinking LLM module interprets a concise textual description of the unknown environment and synthesizes an agent-level exploration strategy, which is then grounded into a sequence of global waypoints through a topological graph. To further improve reasoning efficiency, this module employs a modularity-based pruning mechanism that reduces redundant graph structures. The fast-thinking RL module executes exploration by reacting to local observations while being guided by the LLM-generated global waypoints. The RL policy is additionally shaped by a reward term that encourages adherence to the global waypoints, enabling coherent and robust closed-loop behavior. This architecture decouples semantic reasoning from geometric decision, allowing each module to operate in its appropriate temporal and spatial scale. In challenging simulated environments, our results show that FARE achieves substantial improvements in exploration efficiency over state-of-the-art baselines. We further deploy FARE on hardware and validate it in complex, large scale $200m\times130m$ building environment.

GRATE: a Graph transformer-based deep Reinforcement learning Approach for Time-efficient autonomous robot Exploration

Autonomous robot exploration (ARE) is the process of a robot autonomously navigating and mapping an unknown environment. Recent Reinforcement Learning (RL)-based approaches typically formulate ARE as a sequential decision-making problem defined on a collision-free informative graph. However, these methods often demonstrate limited reasoning ability over graph-structured data. Moreover, due to the insufficient consideration of robot motion, the resulting RL policies are generally optimized to minimize travel distance, while neglecting time efficiency. To overcome these limitations, we propose GRATE, a Deep Reinforcement Learning (DRL)-based approach that leverages a Graph Transformer to effectively capture both local structure patterns and global contextual dependencies of the informative graph, thereby enhancing the model's reasoning capability across the entire environment. In addition, we deploy a Kalman filter to smooth the waypoint outputs, ensuring that the resulting path is kinodynamically feasible for the robot to follow. Experimental results demonstrate that our method exhibits better exploration efficiency (up to 21.5% in distance and 21.3% in time to complete exploration) than state-of-the-art conventional and learning-based baselines in various simulation benchmarks. We also validate our planner in real-world scenarios.