Mobile manipulators such as mechanical loaders and rescue robots consist of a moving base and a one (or multi) degree of freedom arm attached to the base. In many circumstances mobile manipulators are supposed to move on uneven terrains and may get into postures dangerous for their stability. The aim of my work was to develop and implement an active real-time feedback control to be installed on such vehicles to enhance their stability by static and dynamic compensatory motion of their arms. This compensatory motion is similar to what our hands do when we are about to fall down.
Examples of Mobile Manipulators in real life
Classical control methods are based on closed form or approximate analytical solutions which – although very good in simple less degree of freedom robots- in real many degree of freedom mobile manipulators are very slow and non-reliable. For high degree of freedom robots, inherent complexity of equations and the need to real-time answers to the optimization problem sparked us to apply artificial intelligence techniques. We developed an algorithm that invokes combinatory Artificial Neural Networks and Genetic Algorithms to achieve a robust feedback control. We showed the performance of our approach via many numerical examples.
Schematic of problem definition.
– Meghdari A., Naderi, D. and Alam, M.R., “Tipover Stability Estimation for Autonomous Mobile Manipulator Using Neural Network” Proceedings of the 2004 Japan–USA Symposium on Flexible Automation, JUSFA 2004, pp. 19-25.
– Meghdari A., Naderi, D. and Alam, M.R., “Real-Time Compensatory Manipulator Motion Planning for Stabilizing a Mobile Manipulator” Proceedings of the 2003 ASME Int. Design Engineering Technical Conferences, September 2003, pp. 2-6.
– Alam, M.R., Meghdari A. and Naderi, D., “Optimally Stable Mobile Manipulator Path Planning Using Genetic Algorithm” Proceeding of international conference of mechanical engineering (ISME 2003) Tehran-IRAN.