基于激光和单目相机信息融合的智能轮椅避障策略研究

针对智能轮椅使用环境复杂多变,障碍物形状各异,单一传感器无法获得完整的环境信息的问题,提出一种基于激光传感器和单目视觉传感器信息融合的障碍物检测方法。通过单目相机和激光雷达传感器感知智能轮椅周围环境,得到障碍物的形状、距离分布状况等信息；在此基础上提出两种传感器信息的融合策略,建立局部障碍物地图,进一步采用模糊神经网络完成整体避障算法,实现智能轮椅安全、快速避障等功能。实验结果验证了文中所提避障算法的可行性及有效性。

Obstacle Avoidance of Intelligent Wheelchair Based on Laser and Monocular Vision Fusion

As the number of handicapped people increases worldwidely, the role of electric wheelchair becomes important to enhance their mobility. Considering the working environment of intelligent wheelchair is complex and obstacles have different kinds of shapes, a new method based on the monocular vision and laser fusion to obtain the obstacles was proposed. First of all, a camera installed on the intelligent wheelchair was used to detect the information of obstacles such as the shape of obstacles and the distance between the obstacles and the wheelchair. Then, the environment around the intelligent wheelchair was constructed using the information which was gathered by laser radar and the camera. Obstacle avoidance strategies were built by Fuzzy Neural Networks (FNNs). Finally the intelligent wheelchair could realize obstacle avoidance safely and easily by querying the obstacle avoidance strategies. Experimental results verify the effectiveness of the propose method.