Hidden Markov Model Analysis of Force/Torque Information

A new model is developed for prediction and analysis of sensor information recorded during robotic performance of tasks by telemanipulation. The model uses the Hidden Markov Model (stochastic functions of Markov nets; HMM) to describe the task structure, the operator or intelligent controller’s goal structure, and the sensor sig-nals such as forces and torques arising from interaction with the environment. The Markov process portion encodes the task sequence/subgoal structure, and the observation densities associated with each subgoal state encode the expected sensor signals associated with carry-ing out that subgoal. Methodology is described for con-struction of the model parameters based on engineering knowledge of the task. The Viterbi algorithm is used for model based analysis of force signals measured during experimental teleoperation and achieves excellent segmen-tation of the data into subgoal phases. The Baum-Welch algorithm is used to identify the most likely HMM from a given experiment. The HMM achieves a structured, knowledge-based model with explicit uncertainties and mature, optimal identification algorithms. 1