该算法克服了传统BP算法的收敛速度慢，易陷入局部最小点的问题。
The algorithm solved the problems of the conventional BP algorithm such as converging slowly and falling into the local minimum point easily.
与以往的算法相比，所提出的算法可以跳出局部最小点，并使关节收敛到期望构形。
Compared with the past schemes, the proposed one can escape from local minimum points and guarantee the convergence of joint angles to desired configuration.
与神经网络等其它学习方法相比，它的结构通过自动优化的方法计算出来，并且避免了局部最小点、过学习等缺陷。
It has many advantages compared to Article Neural Networks or other learning methods, for example the automatic structure selecting, overcoming the local minimum and over-fitting etc.
应用推荐