Development of remotely operated bomb disposal robot equipped with object weight and poses estimations

The paper presents development of the prototype of remotely operated bomb disposal robot equipped with object weight and pose estimation. The main mission of the prototype of remotely operated bomb disposal robot is how to move the bomb to safe area. The key issue here is how the robot can safely evacuate the bomb with minimum risk. In order to avoid gripped object detached from the grip, the robot needs to know the object’s weight and pose. In this paper, we also present methods for estimating object’s weight and pose. We estimate the object’s weight by utilizing linear regressions and measuring currents flowing into servomotors. The increasing current flow that goes into servomotors are monitored to estimate the object’s weight. We estimate the object’s pose by using several image processing steps such as color conversion from RGB to HSV, color filter, edge detection, and hough transform. The methods have been implemented in proposed system that consists of several processing units (Raspberry Pi 3, Arduino Mega, Arduino Uno), Brushless Electronic Speed Controller, Robot Arm, Video Sender, Analog Multiplexer, 3DR telemetry, Ardupilot Mega with GPS and compass, and SkyZone 3D Googles. We have tested the proposed system under laboratory condition. It succeeded to lift object up to 257.4 gram of weight. Such ability equal to lift small bomb like C-4 or small IED. The robot’s payload could still be possible to be developed while maintaining robot’s dimension. The maximum distance of the robot to be operated remotely is 236 meter. The robot also succeeded to follow the saved waypoints and passed the rock with maximum diameter of 12 cm. The robot can estimate the object weight with accuracy of 27.25% and object pose with average error of 92.3%. By implementing the proposed system under real condition, the risk during bomb evacuation process can be minimized.