In the last years, the target of many researchers has been to create simple robots able to perform coordinated tasks which could not be accomplished by a single robot. This field of robotics is known as swarm, which is based on the observation of social insects. In order to perform this, researchers encounter many problems when developing hardware that will allow a robot capable of simulating the behavior of swarm. The biggest challenge is to provide the robots with autonomy, because the robot must be aware of its battery needs in order to stay alive. The Wanda robot, the robot that we use in this thesis, it has the same problem. This thesis looks foward to extending the life of a single Wanda and, consequently, of a group of Wanda robots. To achieve this we have designed four algorithms, such that the Wanda robots can remain in operation and share a recharging station. An experimental bottom-up approach has been adopted in order to test various strategies to manage collective self-sufficiency, which rely upon low-level mechanisms such as non-direct communication and non-complex decision making. The algorithms designed are tested in DiScoBoTs simulator, that it is adapted for each algorithm. In addition we model a virtual battery that works as the real.
The recent advance of embedded systems has enabled the capability and availability of small mobile robots. The availability of small mobile robots has impact on the possibility of artificial robotic swarms to construct a multirobot system consisting of a large number of (mostly) simple physical robots. In general, the objective of a swarm is to accomplish a task which would otherwise be impossible for an individual simple physical robot. To provide a tool for studying robotic swarms and miniature mobile robots, we have introduced our Wanda robots in 2010. Along with the miniature robots, we have recently also implemented a beamer-assisted arena for robot experiments and a multi-robot- and sensor-simulator. The simulation framework utilizes the graphics hardware to efficiently and effectively simulate the sensor data, including RGB-sensors, camera sensors, infrared communication, and ranging sensors.
As batteries are the main power sources, one non-functional but important issue is the reduction of the power consumption and the improvement of the power availability by recharging. Ideally, with the help of a recharging facility the swarm can continue by shifting the limit from power source to component failure. We have recently introduced the recharging capability in the arena so that the lifetime of the robots can be prolonged. As robots are moving around the arena autonomously, it is essential and critical to have an autonomous algorithm to decide and plan how to utilize the recharging station. In our arena, all the robots share a charging station, which can host several robots at the same time. This makes resource sharing indispensable among individual robots.
This thesis will explore how to achieve a global collaborative and effective use of the shared recharging station. The tasks are as follows: (1) integrate and design the behavior related to the recharging station in the simulator, (2) design a collaborative scheme for recharging planning, and (3) evaluate the performance of the recharging planning. The fundamental knowledge to program with C or C++ is assumed. The tasks can be broken down as follows:
- 30%: analysis
- 50%: implementation
- 20%: evaluations
Any further queries or clarifications can be directed to Msc. Waqaas Munawar, Dipl.-Phys. Alexander Kettler, or Dr. Jian-Jia Chen. If you have an interest in our research activities just drop by our offices for coffee and discussion.