Autonomous Control of a Novel Closed Chain Five Bar Active Suspension via Deep Reinforcement Learning
Nishesh Singh, Sidharth Ramesh, Abhishek Shankar, Jyotishka Duttagupta, Leander Stephen D'Souza, Sanjay Singh
TL;DR
The paper addresses autonomous traversal of rugged planetary terrain while protecting onboard experiments by introducing a novel closed-chain five-bar suspension and an active-control strategy. It employs Soft Actor-Critic (SAC) with PID actuation in a Gazebo/ROS simulation, formulating the problem as a Markov Decision Process with $S=[pitch,roll,distance,height]$ and $A=[a_0,a_1,a_2,a_3]$. SAC optimizes a maximum-entropy objective using two soft Q-functions and a stochastic policy, enabling robust, continuous-action control for obstacle climbing. Results show significant stability gains, with chassis pitch reduced from $21.31^ ext{°}$ to $10.73^ ext{°}$ and a near-constant traversal velocity of $0.7$ m/s, trained in about 4 hours on standard hardware; the work remains simulation-based and points to future integration with path planning and multi-task obstacle avoidance.
Abstract
Planetary exploration requires traversal in environments with rugged terrains. In addition, Mars rovers and other planetary exploration robots often carry sensitive scientific experiments and components onboard, which must be protected from mechanical harm. This paper deals with an active suspension system focused on chassis stabilisation and an efficient traversal method while encountering unavoidable obstacles. Soft Actor-Critic (SAC) was applied along with Proportional Integral Derivative (PID) control to stabilise the chassis and traverse large obstacles at low speeds. The model uses the rover's distance from surrounding obstacles, the height of the obstacle, and the chassis' orientation to actuate the control links of the suspension accurately. Simulations carried out in the Gazebo environment are used to validate the proposed active system.