Intelligent Systems with Applications (Nov 2021)
A bat-pigeon algorithm to crack detection-enabled autonomous vehicle navigation and mapping
Abstract
Autonomous vehicles have nowadays received widespread attention, and path planning is one of the most important components of its autonomous operation. However, due to the long-term use of roads and lack of maintenance, the roads that autonomous vehicles need to pass inevitably have cracks. In this case, while the autonomous vehicles pass through these cracked areas at high speed, it will increase the sense of bumps and even deviate from the originally planned route, which may potentially cause vehicle damage. Therefore, in this paper, we propose an adjustable speed navigation method in light of crack detection for autonomous vehicle path planning, which can automatically adjust the speed in the cracked areas of the road. Based on the obtained image of the road environment, an image processing algorithm to accurately locate the crack is developed. Then, in light of the obtained location of cracks and obstacles, a Bat-Pigeon algorithm (BPA) is proposed to conduct adjustable speed navigation of autonomous vehicles. Take advantage of the same individual speed update rule, we integrate the global search of the Pigeon-inspired optimization (PIO) algorithm and the local search by the Bat algorithm (BA), which can effectively improve the speed and performance of the convergence algorithm. The proposed Bat-Pigeon algorithm navigates the autonomous vehicle to decelerate around small crack areas while planning a collision-free path with the least travel time. In addition, a local reactive navigator is developed that builds environmental maps locally while avoiding dynamic and unknown obstacles. To verify the theoretical advantages of the developed algorithms, we perform comparative experiments under various scenarios. Simulation and comparative studies validate that the proposed algorithms effectively carry out the adjustable speed navigation and mapping of autonomous vehicles based on road conditions in various real-world scenarios.
