Gait generating network for hexapod robot in consideration of fault tolerance

Recently, robots attract attention as a work force in the dangerous and severe environment. Especially, legged robots are extensively researched for the use in such environment. In such environment, it is necessary for the legged robot working to have fault tolerance and adaptability to environment. There are two reasons. The first is that road condition is not maintained by human. The second is that it is difficult to repair robots. In this paper, a gait generating network with fault tolerant, adaptability to irregular terrain and decentralized system are applied to a hexapod robot. In the proposed method, a new network is generated when one of the legs is broken and a new gait is produced with new network which consists of the remaining legs. The method is applied to hexapod robot and experimental simulation is conducted. The generation of gait when walking in slopes and walking with one deficient leg is confirmed.