A Real-Time Obstacle Detection and Reactive Path Planning System for Autonomous Small-Scale Helicopters
Joshua Redding, Jayesh Amin, Jovan D. Bošković, Yeonsik Kang, Karl J. Hedrick, Jason K. Howlett, Scott D. Poll · AIAA Guidance, Navigation, and Control Conference and Exhibit · 2007
This paper presents the design of a comprehensive Multi-layer Architecture for Trajectory Replanning and Intelligent eXecution (MATRIX) system and its evaluation through high-fidelity simulations of the R-50 helicopter in the Real-time Interactive Prototype Technology Integration/Development Environment (RIPTIDE). The main feature of the MATRIX system is that it assures effective real-time obstacle detection & avoidance in a cluttered environment. Specifically, this paper discusses the (i) Design of the MATRIX System; (ii) Design, implementation & testing of individual modules; (iii) Integration & implementation under RIPTIDE; and (iv) Study of the features and limitations of the proposed architecture and algorithms. The MATRIX system consists of several subsystems including: Path Planner based on Rapidly-exploring Random Trees (RRT), shortest path algorithm for pruning of redundant waypoints, a safety corridor for preventing collisions with obstacles, 4D trajectory generator, nonlinear adaptive outer-loop controller, representation of obstacles using Quadtrees, and obstacle detection using a scanning laser and Interacting Multiple Model Kalman Filters (IMM-KF). These subsystems were developed and integrated within RIPTIDE, and extensive testing was carried out using R-50 helicopter dynamics. It was found that the MATRIX system rapidly generates new paths and trajectories to avoid unknown obstacles, is highly robust, and achieves excellent overall performance. I.