Browsing by Author "Md. Masum Billah"

Robots are gradually becoming omnipresent in daily life to take care of everyday chore like vacuuming floor, delivering food in hospital etc. However, these are examples of single robot activity. Difficult tasks like transporting big size objects inside factory floor or mapping of terrain for task planning demands multiple robots work in coordination. Groups of robots can solve problems in fundamentally different ways than individuals while achieving higher levels of performance. However, programming and coordination pose challenges on collaborative activity of all the robots inside a group. This work presents new collaborative algorithm for autonomous multi-agent robot for undertaking reconnaissance mission. A set of communication techniques, sharing behavior to get their position and appropriate attitude which is useful for a group of robots to work together for any kind of reconnaissance mission are developed. A modified gradient communications algorithm is presented here which is flexible to the constantly changing network topology of the multi-agent robots. This provides real-time information that is used to communicate data and to guide robots around the physical environment. Two hexapod robots equipped with inertial sensors like accelerometer and rate gyro are chosen for multi-agent robot navigation in 3D uneven terrain for their advantages over the traditional wheeled counterpart. The sharing behavior and coordination of motion are verified through experiments that built 3D map of a 5x5 m corrugated terrain. Two hexapod robots generated four thousand nodes in half an hour collaborating with each other while travelling over the terrain. Triangulation algorithm is then used for mapping 3D terrain using these nodes. Comparison of the new mapping algorithm with one of the existing algorithms shows that the new algorithm is superior to the existing one. In this research sensory signals are processed in a remotely located wireless connected computer. This poses limitation on the processing speed of data.

Snakes are available all over the world and the flexible biological structure allows snakes to roam over wide range of terrain without facing any difficulty. This especial feature of snake attracted researchers to mimic snake like robots. All the existing snake robots are made of rigid links, which are connected with joints of single degree of freedom. Electrical motors are commonly used as actuators for actuating these joints. Such actuation systems are not suitable for executing versatile snake motion due to their limitations of one degree of freedom. New kind of joints capable of executing two (2D) or three degree (3D) of freedom is a requirement to mimic natural snake locomotion. However, the traditional actuators like electrical motors that are popular in designing snake robots are not able to actuate such 2D or 3D joints. In this research, a new kind of 3D joint has been designed and used to develop a flexible snake robot. As electrical motors are not able to handle this multi-degree of freedom joints, Smart (active) materials in the form of tendons have been used for actuating these new joints. This approach has reduced the complexity of the controller that is developed for executing snake locomotion. A Central Pattern Generator (CPG) based motion control has been implemented for the motion control of the snake robot. The CPG based algorithm was adopted from gait analysis of real snake. The newly developed snake robot is capable of executing 3D motion, and executing online gait transition for obstacle avoidance. The snake robot was tested on three different surfaces possessing three different coefficients of friction. It was found that the maximum speed was achieved with the highest coefficient of friction. The maximum speed of the robot achieved was 2.0 cm per second on the surface of coefficient of friction of 0.46. Head rise of the robot achieved was 4 cm. Though smart materials made the snake robot capable to execute 3D motion, however, due to the limitation of actuation force high speed could not be achieved for the relatively heavy structure of the snake. Head rise of the robot also faced similar limitation. Further study on optimizing size of robot and smart actuator is expected to provide snake robot for real life application.