A Global Optimization Algorithm for 3D Point-Set Registration Based on Quantum Particle Swarm Optimization

Under the background of the rapid development of automatic driving technology, the precise positioning of vehicles has become the key. However, it is difficult to realize precise positioning only with raw sensor data, and point cloud alignment technology is needed. Among them, the iterative closest point (ICP) algorithm, as one of the most popular algorithms for point cloud alignment, tends to fall into the local optimum when the initial position is poor. In this paper, we propose a quantum particle swarm optimization-based ICP optimization strategy to address the local extreme value sensitivity of the ICP algorithm. The quantum particle swarm optimization is used to iteratively update the transformation matrix, and increase the threshold and change the termination conditions to ensure the convergence of the algorithm and achieve high alignment accuracy. Experimental results show that this algorithm significantly improves the alignment accuracy and robustness.