Bio-inspired explainable neuromorphic controller for path tracking and obstacle evasion using an array of Hindmarsh-Rose neurons

This work presents an approach to autonomous navigation utilizing the Hindmarsh-Rose neuron model for integrated path tracking and obstacle avoidance. Inspired by the whisker-mediated touch system in rodents, the proposed methodology combines path following and evasion strategies within a single neural architecture, enhancing adaptability in dynamic environments. The neuron architecture allows for flexible integration with various sensor configurations, reducing the need for extensive re-tuning. The systems performance is demonstrated by adjusting obstacle avoidance and path tracking gains as well as look ahead distances in order to balance stability and responsiveness. As a contribution, this paper presents an alternative algorithm which enables a single structure to control a vehicle for path following and obstacle evasion without relying in blackbox artificial intelligence tools or classic control methodologies.