Survey on model-based biped motion control for humanoid robots

Modeling, analysis and control of robot–object nonsmooth underactuated Lagrangian systems: A tutorial overview and perspectives

“…Hence a controller p n (t k ) is chosen to set a desired q1 (t + k ) with (30), then q2 (t − k ) is deduced from (29), and the final step is to design τ [t k−1 ,t k ] which is τ over [t k , t k+1 ] using (31). Another algorithm is detailed in [3, section 4] which uses flight times and applies to a general class of vibro-impact systems [3,Equation (26)] (it extends the recursive method in [1, section III.B]).…”

“…It involves impulsive multipliers (p n , p t ) considered as arbitrary inputs (impulsive controllability [113]), or bounded multipliers which are functions of time (solutions of the CP in (3) or ( 9) or ( 15)). In reality both are signed nonnegative inputs due to complementarity, see also (29), consequently tools for signed inputs should be used [3]. Notions of controllable twists and wrenches are introduced for manipulation grasping [16], where ranges and null space of H 1 (q) and H 2 (q) are the basic analysis tools.…”

“…walking and running robots [4,25,26,27,28,29,30,31,32], jumping [33], climbing [34], crawling robots [35,36,37] (where the center of gravity plays the role of the object), kinematic chains with joint clearance [3], backlash effects and compensation in machine tools, see early studies in [38,39,40,41,42,43,44,45],…”

“…Remark 2. Several classes of nonsmooth robotic systems (bipedal locomotion [4,25,26,27,28,29,30,71,72], manipulation [16,17,24,73,74,75,76,9,77], systems with joint clearance [78,79], hopping robots [33], pushing tasks [80], quadruped robots [81,82], snake robots [36], cable-driven manipulators [46,47], spherical robots with internal rotors [83]) have already been the object of survey articles in the Automatic Control or in the Robotics literature. It is therefore outside the scope of this article, to survey them exhaustively once again, since this would yield repetition and far too many references (probably several thousands).…”

Robust Walking for Humanoid Robot Based on Divergent Component of Motion

“…They measured changes in the acceleration of the robot’s CoM and mapped changes in ZMP onto the new support area based on variations. However, the conventional LIPM has strict limits on the height and angular momentum of the CoM, which greatly limit further robot application [ 20 , 21 , 22 ]. To solve this problem, scholars have improved the LIPM by considering the changes in height and angular momentum of the CoM.…”