Maximum acceptable communication delay for the realization of telesurgery

Abstract

Aim: To determine acceptable limits of communication delays in telesurgery, we investigated the impact of communication delays under a dynamic environment using a surgical assist robot. Previous studies have evaluated acceptable delays under static environments. Effects of delays may be enhanced in dynamic environments, but studies have not yet focused on this point.

Methods: Thirty-four subjects with different surgical experience (Group1: no surgical experience; Group2: only laparoscopic surgical experience; Group3: robotic surgery experience) performed 4 tasks under different delays (0, 70, 100, 150, 200, or 300 ms) using a surgical assist robot. Task accomplishment time and total movement distance of forceps were recorded and compared under different communication delays of 0-300 ms. In addition, surgical performance was compared between Group1or Group2 without delay and Group3 with communication delays.

Results: Significant differences in task accomplishment time were found between delays of 0 and 70 ms, but not between delays of 70 and 100 ms. Thereafter, the greater the communication delay, the longer the task accomplishment time. Similar results were obtained in total movement distance of forceps. Comparisons between Group3 with delay and Group1 or Group2 without delay demonstrated that surgical performance in Group3 with delay was superior or equal to that of Group1 or Group2 without delay as long as the delay was 100 ms or less.

Conclusions: Communication delays in telesurgery may be acceptable if 100 ms or less. Experienced surgeons with more than 100 ms of delay could outperform less-experienced surgeons without delay.

Fig 1. Actual protocol.

A total of 4 sets were performed, with the last 2 sets performed on different days. * Each task was performed 9 times with different delay settings, as follows: Delay 0 ms → Delay 0 ms → Delay 0 ms → Delay 70 ms → Delay 100 ms → Delay 150 ms → Delay 200 ms → Delay 300 ms → Delay 0 ms.

Fig 2. Details of the system used in this experiment.

The minimum communication delay was 20 ms, considering the typical performance of domestic communications. The minimum total communication delay was thus 70 ms, including the decoding and encoding times. T_N = {0,20,50,100,150,250}ms, T_E+T_D≈50ms, Delay defined in this paper T is: T = T_N+T_E+T_D = {0,70,100,150,200,300}ms, *Encoder/decoder unused when T_N = 0.

Fig 3. Boxplot of task completion time of each task by all participants.

Fig 4. Task completion time for each Group under each delay.

* shows significant difference (P value<0.05). Values are median [minimum-maximum].

Fig 5. Boxplot of movement distance of forceps of each task by all participants.

Fig 6. Total movement distance of forceps for each group under each delay.

* shows significant difference (P value<0.05). Values are median [minimum-maximum].

Fig 7. Boxplot of comparing task completion time between Group 3 with delay and Group 1 or Group2 without delay.

G1(0), Group 1 without delay; G2(0), Group 2 without delay; G3(0), Group 3 without delay; G3 70, Group 3 with delay of 70 ms; G3 100, Group 3 with delay of 100 ms; G3 150, Group 3 with delay of 150 ms; G3 200, Group 3 with delay of 200 ms; and G3 300, Group 3 with delay of 300 ms.