CN-CBF: Composite Neural Control Barrier Function for Safe Robot Navigation in Dynamic Environments

TL;DR

A simple yet effective neural CBF design method for safe robot navigation in dynamic environments by employing the idea of a composite CBF, where multiple neural CBFs are combined into a single CBF.

Abstract

Safe navigation of autonomous robots remains one of the core challenges in the field, especially in dynamic and uncertain environments. One of the prevalent approaches is safety filtering based on control barrier functions (CBFs), which are easy to deploy but difficult to design. Motivated by the shortcomings of existing learning- and model-based methods, we propose a simple yet effective neural CBF design method for safe robot navigation in dynamic environments. We employ the idea of a composite CBF, where multiple neural CBFs are combined into a single CBF. The individual CBFs are trained via the Hamilton-Jacobi reachability framework to approximate the optimal safe set for single moving obstacles. Additionally, we use the residual neural architecture, which guarantees that the estimated safe set does not intersect with the corresponding failure set. The method is extensively evaluated in simulation experiments for a ground robot and a quadrotor, comparing it against several baseline methods. The results show improved success rates of up to 18\% compared to the best baseline, without increasing the conservativeness of the motion. Also, the method is demonstrated in hardware experiments for both types of robots.

Figures (8)

• Figure 1: CN-CBF structure.
• Figure 2: Block diagram of the proposed CN-CBF safety filter.
• Figure 3: Slice of the learned neural CBF for relative dynamics in the case of the ground robot ($\theta_{rel} = \pi; v_o = 1.2$).
• Figure 4: Results from the simulation experiments with the ground robot model.
• Figure 5: An example scene from the simulation experiments with the ground robot model, which includes the 3D simulator (left), visualization of the local perception (middle), and the corresponding slice of the CN-CBF backprojected to the world frame (right).