SwarmPRM: Probabilistic Roadmap Motion Planning for Large-Scale Swarm Robotic Systems

Yunze Hu,Xuru Yang,Kangjie Zhou,Qinghang Liu,Kang Ding,Han Gao,Pingping Zhu,Chang Liu

2024-10-13

Abstract:Large-scale swarm robotic systems consisting of numerous cooperative agents show considerable promise for performing autonomous tasks across various sectors. Nonetheless, traditional motion planning approaches often face a trade-off between scalability and solution quality due to the exponential growth of the joint state space of robots. In response, this work proposes SwarmPRM, a hierarchical, scalable, computationally efficient, and risk-aware sampling-based motion planning approach for large-scale swarm robots. SwarmPRM utilizes a Gaussian Mixture Model (GMM) to represent the swarm's macroscopic state and constructs a Probabilistic Roadmap in Gaussian space, referred to as the Gaussian roadmap, to generate a transport trajectory of GMM. This trajectory is then followed by each robot at the microscopic stage. To enhance trajectory safety, SwarmPRM incorporates the conditional value-at-risk (CVaR) in the collision checking process to impart the property of risk awareness to the constructed Gaussian roadmap. SwarmPRM then crafts a linear programming formulation to compute the optimal GMM transport trajectory within this roadmap. Extensive simulations demonstrate that SwarmPRM outperforms state-of-the-art methods in computational efficiency, scalability, and trajectory quality while offering the capability to adjust the risk tolerance of generated trajectories.

Robotics

What problem does this paper attempt to address?

This paper aims to solve the motion planning problem in large - scale robot swarm systems. Specifically, traditional motion planning methods often face a trade - off between scalability and solution quality when dealing with large - scale robot swarms, because as the number of robots increases, the joint state space grows exponentially, leading to a sharp increase in computational complexity. To solve this problem, this paper proposes a hierarchical, scalable, computationally efficient and risk - aware sampling - based motion planning method named SwarmPRM.

### Main Problems
1. **Trade - off between Scalability and Solution Quality**:
- When dealing with large - scale robot swarms, traditional motion planning methods are difficult to ensure scalability and solution quality simultaneously due to the exponential growth of the joint state space.
- For example, centralized multi - agent motion planning methods (such as RRT) are computationally infeasible when dealing with large - scale swarms.

2. **Challenges in Collision Detection**:
- In large - scale robot swarms, traditional collision detection methods are difficult to effectively handle uncertainty and risk, especially collision avoidance in high - density environments.

### Solutions
1. **Hierarchical Motion Planning Strategy**:
- **Macro - stage**: Represent the entire robot swarm as a Gaussian Mixture Model (GMM), and construct a probability map (Gaussian Roadmap) in the Gaussian space to generate the transfer trajectory of the GMM.
- **Micro - stage**: Each robot tracks the macro - state through a local tracking control law while avoiding collisions with other robots and obstacles.

2. **Risk - aware Collision Detection**:
- Introduce the Conditional Value - at - Risk (CVaR) as the risk measure for collision detection to ensure that the generated trajectories are risk - aware.
- In the specific implementation, calculate the signed distance function (SDF) distribution between samples and obstacles and constrain its CVaR value within the safety threshold.

3. **Optimized Trajectory Generation**:
- Use the Linear Programming (LP) method to calculate the optimal GMM transfer trajectory, thereby finding the lowest - cost path on the generated roadmap.

### Main Contributions
1. **Sampling in Gaussian Parameter Space**:
- Sample in the Gaussian parameter space to construct a Gaussian probability map, where each node represents a different Gaussian distribution, and the cost and geodesic path between nodes are determined based on the Wasserstein metric.

2. **CVaR - based Collision Detection**:
- Develop a systematic collision detection method that uses CVaR as a risk measure to perform collision detection when constructing the Gaussian probability map.

3. **Linear Programming - optimized Trajectory**:
- Design a linear programming formula to calculate the optimal macro - GMM trajectory on the constructed roadmap, and verify the superiority of SwarmPRM in terms of computational efficiency, scalability and trajectory quality through extensive simulations.

### Summary
SwarmPRM effectively solves the motion planning problem of large - scale robot swarm systems by combining hierarchical strategies and sampling - based methods, achieving the unification of scalability, computational efficiency, risk - awareness and trajectory flexibility.