Southwest Research Institute (SwRI) has developed a motion prediction system to help autonomous vehicles predict where pedestrians will go.
The computer vision tool uses a deep learning algorithm to predict motion by observing biomechanical movements from the pelvic area. Technologies have been able to track and predict movements in a straight line but struggled with sudden changes. Researchers optimized temporal convolutional network algorithms, helping it predict sudden changes in motion within milliseconds.
The temporal design uses a convolutional neural network to process video data, optimizing dilation in network layers to learn and predict trends at a higher level.
The research team used SwRI’s markerless motion capture system, which automates biomechanical analysis in sports science. Using camera vision and perception algorithms, the system provides insights into kinematics and joint movement.