Efficiency of Cooperation between Human and Remote Robot System with Force Feedback

Transcription

1 Efficiency of Cooperation between Human and Remote Robot System with Force Feedback Yuichi Toyoda, Pingguo Huang, Yutaka Ishibashi, Yuichiro Tateiwa, and Hitoshi Watanabe * Nagoya Institute of Technology Seijoh University * Tokyo University of Science IEEE ISR, 2018/08/25, Shenyang China

2 Outline Background Purpose Remote Robot System with Force Feedback Experiment Method Experiment Results Conclusion

3 Background (1/2) Remote robot systems with force feedback have been actively researched Especially the cooperative work in which multiple robots collaboratively move an object by holding the object together or one robot moves and hands over an object to another robot Force feedback Humans can feel the shape, softness and weight of a remote object Accuracy of remote cooperative work are expected to be improved largely

4 Background (2/2) When the force information is transferred over a quality of service (QoS) non-guaranteed network like the Internet Network delay, delay jitter, and packet loss unstability phenomena such as vibrations of the robot and device may occur Quality of Experience (QoE) may seriously deteriorate QoS control, Stabilization control In order to achieve efficient control, we need to clarify the influences of network delay, delay jitter, and packet loss

5 Purpose(1/2) *1 Taguchi et al., IEICE Global Conference, B-11-17, Mar Previous study *1 Work in which two industrial robots move an object collaboratively by using two remote robot systems is dealt with the influence of network delay on the collaborative work is investigated Problem only cooperative work in which two industrial robots move an object by holding the object together is dealt with It is important to handle the cooperative work in which one human (or industrial robot) moves an object and hands over the object to another human (or industrial robot), who receives it

6 Purpose(2/2) This work Before dealing with cooperative work in which one robot moves an object and hands the object to another robot, we handle cooperative work between a human and a remote robot system with force feedback investigate the influences of network delay on the work efficiency by experiment

7 Configuration of remote robot system with force feedback Haptic interface device: Geomagic Touch Industrial robot: Mitsubishi RV-2FB-Q Protocol: UDP Video resolution: pixels

8 Force calculation *2 K. Suzuki et al., Proc. IEEE GCCE, Oct The reaction force F (m) t applied to the haptic interface device at time t ( 1) is calculated as follows *2 F : Force received from the slave terminal at time t K scale : Force scale When F t (m) exceeds the maximum force (3.3 N) applied to the haptic interface device, the reaction force is set to 3.3 N K scale is set to 0.1 by a preliminary experiment

9 Calculation for position of Industrial robot arm *2 K. Suzuki et al., Proc. IEEE GCCE, Oct S : Position vector of the industrial robot at time t ( 2) M : Position vector of haptic interface device received from the master terminal at time t ( 2) V (=S S ): Velocity of the industrial robot at time t ( 2) V : The maximum value of the velocity (5mm/s)

10 Experiment method(1/2) Deal with two types of work work 1: a human moves and hands over a wooden block to the industrial robot, and the industrial robot receives the block by grasping it with the electric hand attached to the industrial robot arm work 2: the human receives the wooden block grasped by the electric hand The work are done in two directions direction 1: the magic hand is in front of the electric hand attached to the industrial robot direction 2: the magic hand is at 45 degrees to the left in front of the electric hand

11 Experiment method(2/2) The initial position of the electric hand was set at the same height as the magic hand The distance between the electric hand and magic hand was set to 5 cm The additional delay is changed from 0 ms to 200 ms at intervals of 50 ms We carried out the experiment 10 times for each combination of work contents, The average time from the moment the work is started until the instant the human or the industrial robot receives the block receiving/handing over, and the additional delay The combinations were selected in random order. The average operation time is measured

12 Video clip of experiment (1/2) Work 1: the industrial robot receives the block by grasping it with the electric hand attached to the industrial robot arm Direction: the magic hand is at 45 degrees to the left in front of the electric hand Network delay: 0ms

13 Video clip of experiment (2/2) Work 2: a human receives the wooden block grasped by the electric hand Direction: the magic hand is at 45 degrees to the left in front of the electric hand Network delay: 0ms

14 Experiment results (1/3)

15 Displayed image (45 degrees to the left) Hard to recognize the position of the industrial robot arm, especially the positional relation between the electric hand and the block

16 Experiment results (1/3)

17 Video clip of experiment (45 degrees to the left, 200ms) Work 1: the industrial robot receives the block by grasping it with the electric hand attached to the industrial robot arm Direction: the magic hand is at 45 degrees to the left in front of the electric hand Network delay: 200ms

18 Experiment results (2/3) Work 1: receives the block from a human Direction 1: the magic hand is in front of the electric hand Network delay: 0ms x axis

19 Experiment results (3/3) Work 2: hands over the block to the human Direction 2: the magic hand is in front of the electric hand Network delay: 0ms x axis

20 Conclusion This paper investigated the influences of network delay on efficiency of cooperative work between a human and a remote robot system with force feedback The average operation time of work increases as the network delay becomes larger When the electric hand receives the block from the magic hand, the force of haptic interface device and force sensor of industrial robot are larger than those when the electric hand hands over the block to the magic hand

21 Future work Investigate the influence of network delay by carrying out the cooperative work among multiple industrial robots Handle other types of cooperative work such as work conveying an object held together by two electric hands