A Vibrotactile Feedback Based Cooperative Navigation System For The Blind

With the development of technology,the researchers have designed many smart electronic travel aids for the blind.However,these devices have some defects and there is some room for improvement,such as wearable electronic travel aids increase the burden of the blind users and in traditional navigation systems,there are no fixed formations between robots and the users,and users have to follow the robots passively.In view of these drawbacks,this paper utilizes a mobile robot to guide the blind users and proposes a leaderfollower control model based human-robot formation control strategy.In this system,the blind users can freely select his/her walking speed.In order to maintain the distance and relative bearing between the robot and the user,the linear and angular velocity of the robot is regulated based on user's linear velocity,and the user is prompted to adjust his/her orientation based on his/her relative position to the robot.Finally,the blind users are guided to the goal successfully.This paper designs a vibrotactile feedback based cooperative navigation system for the blind.The guide robot is composed of detecting module,lower control module and a wrist worn vibrotactile interface.The detecting module comprises a laser range radar,an inertial measurement unit and a Kinect sensor.The laser range radar collects the environment information,which is used to achieve simultaneous localization and mapping.The inertial measurement unit is used to get angular velocity of the robot.The Kinect sensor is used to achieve human tracking.The lower control module is responsible for receiving the instructions from the upper control module and then translating the instructions to parameters which are sent to actuators.The wrist worn vibrotactile interface maps the guide instruction to the vibration motors,which provides vibration hints for the blind user.On the base of Robot Operating System,this paper achieves 2D laser simultaneous localization and mapping and builds up the environment map.To enable the blind user to walk freely,this paper proposes a human-centered path planning algorithm which is based on the dynamic window approach.Based on the leader-follower formation control model in multi-robot field,this paper proposes a human-robot formation control model.The blind user's position information relative to the robot is obtained through the Kinect sensor.Depending on the blind user's position,the corresponding guide instruction is sent to the blind user by the wrist worn vibrotactile interface.In this way,the relative distance and relative bearing between the robot and the blind user can be maintained.Finally,the blind user will be guided to the destination successfully.Several experiments have been conducted on the proposed system.The vibrotactile feedback experiment validates the effectiveness of the designed wrist worn vibrotactile interface.The robot automatic navigation experiment tests the ability of the robot to build up maps and navigate automatically.The results show that the average deviation from the final robot position after automatic navigation to the goal target is 7cm,which can meet the demand of indoor navigation for the blind.The Kinect based human tracking experiment tests the human tracking performance of the Kinect.The result shows that the best human tracking performance can be achieved when the distance between the blind user and the Kinect is 1.5m.The human-robot cooperative navigation system performance experiment tests the guide performance of the proposed system.The results show that the average formation error of all testers is 0.18 m,smaller than the average formation error of the similar navigation system 0.23 m,which demonstrates the effectiveness of the proposed system.