2017/12/22 by Luiza Mici, Mici, Luiza, German I. Parisi +3
Engineering · Neuroscience · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Motor Control and Adaptation #Muscle activation and electromyography studies #Neural and Evolutionary Computing (cs.NE) #Robot Manipulation and Learning #Robotics (cs.RO)
paper · pdf · doi:10.48550/arxiv.1712.08521
openalex publication_date 2017/12/22 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
During visuomotor tasks, robots must compensate for temporal delays inherent\nin their sensorimotor processing systems. Delay compensation becomes crucial in\na dynamic environment where the visual input is constantly changing, e.g.,\nduring the interacting with a human demonstrator. For this purpose, the robot\nmust be equipped with a prediction mechanism for using the acquired perceptual\nexperience to estimate possible future motor commands. In this paper, we\npresent a novel neural network architecture that learns prototypical visuomotor\nrepresentations and provides reliable predictions on the basis of the visual\ninput. These predictions are used to compensate for the delayed motor behavior\nin an online manner. We investigate the performance of our method with a set of\nexperiments comprising a humanoid robot that has to learn and generate visually\nperceived arm motion trajectories. We evaluate the accuracy in terms of mean\nprediction error and analyze the response of the network to novel movement\ndemonstrations. Additionally, we report experiments with incomplete data\nsequences, showing the robustness of the proposed architecture in the case of a\nnoisy and faulty visual sensor.\n