仿真假体视觉下躲避静态障碍的任务研究

Task research on avoiding static obstacles under simulated prosthetic vision

为了探讨在虚拟场景中仿真假体视觉下躲避静态障碍时的情况,文中利用Unity游戏引擎搭建静态障碍物的虚拟场景,通过着色器(Shader)屏幕后处理进行像素化,在分辨率为48×48像素情况下,利用被试操作键盘和鼠标控制软件内的玩家(Player)完成相应避障任务。选取25名被试者分别进行像素点缺失、像素点大小、高对比度等一系列的避障实验,任务过程中记录任务完成时间、跳失误、碰撞、掉出等,并对数据进行分析。结果表明,所有失误中跳失误发生次数最多,其次是碰撞。在正常条件下,提高对比度和像素点大小对减少失误有帮助;在像素点缺失的情况下,提高像素点的大小能够帮助被试躲避障碍;同时提高对比度和像素点大小避障效果最好。被试在假体视觉的环境中不能准确判断自己到障碍物之间的距离,距离过远或过近都会造成避障失误,可通过提高障碍物和背景的对比度以及光幻视点大小减少失误。文中研究可为人工视觉假体系统的设计和开发提供前期的理论储备。

In order to explore the situation of avoiding static obstacles in the virtual scene,Unity game engine is used to build the virtual scene of static obstacles,the screen post-processing of shader is conducted to pixelated,and at the resolution of 48 × 48 pixels,the tested operation keyboard and mouse are used to control the Player in the software to complete the corresponding obstacle avoidance task. A total of 25 subjects are selected to conduct a series of obstacle avoidance experiments(including missing pixels,pixel size,high contrast,etc.). In the process of the task,the completion time,jumping error,collision,falling out and so on are recorded,and the data is analyzed. The results show that the number of jumping errors is the most,followed by collision. Under normal conditions,improving the contrast and pixel size is helpful to reduce errors,but in the case of missing pixels,improving the pixel size can help the subjects avoid obstacles and has the best effect on improving the contrast and pixel size at the same time. In the prosthetic vision environment,the subjects cannot accurately judge the distance between themselves and the obstacles. If the distance is too far or too close,it will cause obstacle avoidance errors.Errors can be reduced by improving the contrast between the obstacles and the background and the size of the light illusion viewpoint. This study can provide a theoretical basis for the design and development of artificial vision prosthesis system.