Improved collision detection system inspired from the neural network of the locust

This paper presents a new improved system for collision detection that can be used in guiding visually impaired people. The system is inspired by the LGMD (Lobula Giant Motion Detector) neuron found in the neural system of the locust, which can avoid collisions by using visual information. The original algorithm used for collision detection was improved using different median filters. The new algorithm was implemented and simulated in MATLAB. Also a hardware system is being designed for collision detection using this algorithm. The system was confirmed to respond selectively to colliding object even in complicated real-world situations.