Towards Obstacle-Avoiding Control of Planar Snake Robots Exploring Neuro-Evolution of Augmenting Topologies

TL;DR

The paper tackles obstacle avoidance for planar snake robots in cluttered environments by evolving a neural controller with NEAT that outputs two serpenoid gait parameters, $ω$ and $φ_0$. The network receives LiDAR scans and pose information, and its topology and weights co-evolve to maximize a reward balancing progress to the goal against collisions, within a PyBullet simulation. Empirical results show NEAT outperforms several state-of-the-art methods in routing time and collision count while using a compact network (≈19 KB with 4,897 parameters), and ablation reveals a 45-degree head-angle as optimal. The work demonstrates a resource-efficient, topology-evolving approach for adaptive serpentine locomotion in obstacle-dense scenarios, with potential for real-time deployment and extensions to other NEAT variants.

Abstract

This work aims to develop a resource-efficient solution for obstacle-avoiding tracking control of a planar snake robot in a densely cluttered environment with obstacles. Particularly, Neuro-Evolution of Augmenting Topologies (NEAT) has been employed to generate dynamic gait parameters for the serpenoid gait function, which is implemented on the joint angles of the snake robot, thus controlling the robot on a desired dynamic path. NEAT is a single neural-network based evolutionary algorithm that is known to work extremely well when the input layer is of significantly higher dimension and the output layer is of a smaller size. For the planar snake robot, the input layer consists of the joint angles, link positions, head link position as well as obstacle positions in the vicinity. However, the output layer consists of only the frequency and offset angle of the serpenoid gait that control the speed and heading of the robot, respectively. Obstacle data from a LiDAR and the robot data from various sensors, along with the location of the end goal and time, are employed to parametrize a reward function that is maximized over iterations by selective propagation of superior neural networks. The implementation and experimental results showcase that the proposed approach is computationally efficient, especially for large environments with many obstacles. The proposed framework has been verified through a physics engine simulation study on PyBullet. The approach shows superior results to existing state-of-the-art methodologies and comparable results to the very recent CBRL approach with significantly lower computational overhead. The video of the simulation can be found here: https://sites.google.com/view/neatsnakerobot

Figures (6)

• Figure 1: Snake Robot Schematic
• Figure 2: Obstacle Dense Environment: The blue cylinders represent the obstacles, the white blocks are the arena walls, and the green cylinder represents the goal beacon. The red dot is the starting point of the snake
• Figure 3: Comparison of the path taken by the snake in NEAT, CBRL Jia2021, MFRL 8103164 and DLDC Sartoretti_2019. The cross is the starting point, and the dot represents the end point of the snake.
• Figure 4: Performance metrics of the NEAT controller. The plots show (top row, left to right) the evolution of the x-position, y-position for NEAT, CBRL Jia2021, MFRL 8103164, DLDC Sartoretti_2019, and head angle of NEAT over time, and (bottom row, left to right) control frequency $\omega$, offset $\phi_0$, and distance to goal with time of the NEAT controller.
• Figure 5: Measured torques of all joints of the snake.