Title |
Development of Modular DEAS (mDEAS) and its Application to Optimal Trajectory Generation of Biped Walking |
Authors |
김은숙(Kim, Eun-Su) ; 김조환(Kim, Jo-Hwan) ; 김종욱(Kim, Jong-Wook) |
Keywords |
Numerical optimization ; DEAS ; Heuristics ; Humanoid ; Biped walking |
Abstract |
This paper newly proposes a modular type dynamic encoding algorithm for searches (DEAS) which partitions the whole parameters into several modules and carries out exhaustive DEAS for each module. uDEAS is used to measure parameter sensitivities to the cost function, and the variables whose sensitivities are similar are grouped to make a module. The proposed optimization method is applied to optimal trajectory generation for biped walking of a humanoid. and the optimization result is compared with those of the former versions of DEAS. |