Gesture Recognition Based on Localist Attractor Networks with Application to Robot Control [Application Notes]

In this work, we proposed an online gesture recognition method based on LAN (Localist Attractor Networks). It was employed to recognize human gestures using streams of feature vectors extracted from real-time sensory data. As an application, the gesture recognition system was used to instruct the robot to execute the predefined commands such as moving in different directions, changing speed, stopping and so on. Experimental results showed a high accuracy in controlling the mobile robot using gesture recognition. This system provides a flexible and easy-to-use human-robot interface to control a robot. The user only needs to demonstrate all gesture patterns a few times before starting the control tasks. The robot is able to memorize all patterns and recognize a given gesture. Furthermore, in order to define a new pattern corresponding to a new control task, users only need to demonstrate this pattern. It is advantageous in the areas of service robots since many of them will be operated by non-expert users.