摘要
软体机器人具有优越的柔软性能,能够灵活的穿越狭小的空间,并且对非结构化环境具有较强的自主适应能力.驱动方式和路径规划是软体机器人的关键,其驱动分为有缆驱动和无缆驱动,采用气动、形状记忆合金、电活性聚合物、聚合凝胶等作为驱动器.气动、形状记忆合金之类的驱动器灵活度低、自由度少;电活性聚合物以及聚合凝胶之类的驱动器灵活度高、自由度高.软体机器人的路径规划主要采用人工智能算法,在实际使用中还存在一系列的问题需要继续研究.比如概率路线法和碰撞检测法都易陷入局部最小点与最优点;遗传算法运算效率不高、在线规划困难;神经网络算法泛化能力差等.现在可用的智能算法都只适用特定的物体而不适用通用可变形物体,未来需要致力于柔性驱动器以及新型路径规划算法的研究.
Characterized by its amazing softness, Soft robot can agilely get through cramped space and has strong ability of self-adaption to unstructured environment. Driving mode and path planning are the keys of soft robot. Its drive mode can be divided into cable driven and wireless drive. The materials used in actuators include pneumatic, shape memory alloys, electroactive polymers and polymer gel etc. Pneumatic and shape memory alloys have low flexibility and low degrees of freedom; while materials, like electroactive polymers and polymeric gel, have high flexibility and high degree of freedom. When the path planning mainly adopts artificial intelligence algorithms, many problems in application are still left to be further researched. Probability path method and impact test are easily trapped in the local minimum point and the most advantage; Genetic algorithm, claimed as easy to be combined with other algorithms, is low in operation efficiency and difficult in online programming and its evolutionary effect is not obvious; Neural network enjoys the reputation of high learning ability, simple calculation, fast convergence speed, but it has poor generalization ability. Now the available intelligent algorithm is only applicable to the particular object and cannot be applied to general deformable objects. So it is necessary to focus on the flexible software robots drive and new path planning algorithm in the future study.
出处
《南通大学学报（自然科学版）》
CAS
2013年第3期28-33,共6页
Journal of Nantong University(Natural Science Edition)
基金
国家自然科学基金项目(61105111
61005054)
作者简介
尤小丹（1988-），女，硕士研究生．
通信联系人：陈峰（1977-），男，副教授，博士，主要研究方向为机器人和信息获取．E—mail：hardwk@163．com
