Integration of intelligent agents in an adaptive aiding system for the blind.

摘要

Adaptive human-machine aiding is a concept that involves using aiding/automation only at those points in time when human performance needs support to meet operational requirements. The concept involves two major principles: (1) Adaptation, (2) Interaction. This concept is implemented and tested in a new travel aid for the blind--the Navbelt. The device consists of a portable computer, ultrasonic sensors and stereophonic headphones. The computer applies navigation and obstacle avoidance technologies that were developed originally for mobile robots. The computer then uses stereophonic imaging techniques to process the signals from the sensors and relays them to the user by stereophonic headphones. The imaging techniques convey information to the user that can be interpreted as an acoustic "picture" of the surroundings (image mode), or as the recommended travel direction (guidance mode). Providing information about the recommended travel direction is a breakthrough in travel aids for the blind. The work done in this research includes three phases: (1) Engineering development of the Navbelt. (2) Development of theoretical propositions related to adaptive aiding systems. (3) Experiments with the Navbelt to test these propositions. In the first phase an experimental prototype and a simulator were developed. The simulator was used to investigate the effect of different auditory signals on human performance, and also for training new users. In the second phase three theoretical propositions were developed. These propositions deal with the problem of sharing control and allocating tasks between human and machine in adaptive aiding systems, adaptive information transfer of auditory signals to a blind traveler in order to optimize the information flow, and real time mobility assessment of a human traveling with the Navbelt. The third stage included experiments with the Navbelt prototype and simulator in order to test the theoretical propositions. The results of these experiments provided important data for modifying the engineering design, as well as clarifying theoretical issues. In these experiments subjects were able to travel safely and comfortably in unfamiliar environments at an average speed of 0.8 m/sec in the guidance mode and 0.5 m/sec in the image mode.