The Task Of Avoiding Obstacles Under Simulated Prosthetic Vision

Visual prosthesis can bring light to the patients who have been blind for many years,but the sight mentioned here is different from that of normal people have.Under prosthesis vision,the phosphenes in the field of viewpoint was limited,and there was no hue,only the black-andwhite phosphenes.Phosphenes loss and distortion are one of the conditions to make matters worse.To form prosthesis vision,a micro-electrode array is implant on the visual pathway can stimulate certain areas with a mild electric current.In this study,a virtual reality obstacle avoidance tasks with Unity engine was built,and Unity Shader was used to make a prosthesis vision.The formal experiment was divided into static obstacle avoidance test task and dynamic obstacle avoidance task.The results showed that the final completion time increases with the increase of dropout,and the total number of mistakes also increases.It was not obvious to improve the contrast to reduce the test time,but it was counterproductive under the condition of dropout.With the increase of the size of the phosphenes,the test time was significantly reduced,while the test time is significantly reduced with the increase of the phosphenes and the contrast.However,the retrial time was slightly increased.Analysis of the number of mistakes can be concluded: low obstacles were the main cause of making mistakes.The proportion of all test jumping mistakes was the largest.Increasing of the phosphenes and the contrast can significantly reduce the occurrence of jump mistakes.The results of dynamic obstacle avoidance task under simulated prosthesis vision were as follows: the retrial time of dynamic obstacle avoidance experiment was relatively less than that of static obstacle avoidance experiment.There was little change in time under various conditions.Increasing the phosphenes and contrast has little effect on the dynamic obstacle task time.So in the reality world,the prosthesis wearer should take more practice for perception training to improve the ability of obstacle avoidance and speed perception.For vertical obstacles,the order of the number of errors from high to low was: vertical fast,vertical slow,and vertical.The vertical moving obstacle speed was too fast or too slow would produce mistakes,a moderate speed can effectively reduce errors.The total number of mistakes of horizontal obstacles was lower than that of vertical obstacles,and the threat to subjects was not as great as that of vertical obstacles.Compared with the performance data of static obstacle avoidance task and dynamic obstacle avoidance task,the following conclusions were drawn: the total number of mistakes of dynamic obstacle avoidance was higher than that of static obstacle avoidance,and it was suggested that the prosthesis wearer should increase the time for speed perception training in real life.At the same time,improving the contrast and increasing the pixels have a great difference in the number of mistakes between dynamic and static obstacles.This condition can effectively reduce the number of mistakes in static obstacle avoidance,but has little effect on the number of mistakes in dynamic obstacle avoidance.The results in this study can provide idea for the development of image processing,and a guiding scheme for the rehabilitation training of prosthesis wearers.