Please use this identifier to cite or link to this item:
|Title:||Automatic docking and charging station for a mobile robot platform||Authors:||Lim, Shaune Ji Zheng||Keywords:||DRNTU::Engineering::Mechanical engineering::Mechatronics
|Issue Date:||2015||Abstract:||In a technologically advanced society, there is an increase in the usage of robots to replace humans in performing tasks. Robots can work without rest, which increases the productivity of the tasks. In order to do so, the robot has to have a constant stream of power supply to allow its continuous operation. This is achievable if the robot is stationary and can obtain power from the wall socket. However for the operation of a mobile robot, a battery would have to be utilised. Charging of the battery would usually be achieved through manual means. However for a fully autonomous robot, autonomy of this operation is preferred. This project is part of a group that establishes MAVEN, a Mobile Avatar for Virtual Engagement by NTU. MAVEN is designed to be a fully autonomous telepresence and museum tour guide robot. This project collaborates with two other modules: a navigation module and an automated speech dialogue management module, integrating with each other to form a mobile platform capable of guiding visitors around a museum environment. The objective of this project is to design and develop a docking station for MAVEN, as well as the charging subsystem and the algorithm for the docking operation. The docking module plays an important part for MAVEN’s autonomous operations. It allows for MAVEN to dock autonomously during operation to replenish battery levels without the help of humans. A unique docking station was designed and manufactured for this application, following the design process of research, conceptual design, computer aided design and fabrication. The design process encompasses of the principles and rules of embodiment design. The prototype docking station consists of four main parts. Firstly, a housing is designed to hold the charger. The second part consists of a male alignment plate attached to a two degrees-of-freedom mechanism of a revolute and a prismatic joint. A pair of springs are attached to either sides of the alignment mechanism to allow for the position of alignment plates to reset itself when MAVEN undocks. Next, the stock charger is attached at a height and position corresponding to MAVEN’s charging port. The stock charger is used in order to remove the need of modifying the design of MAVEN. Lastly, a marker is attached to a backing plate on the docking station for the camera to capture the position of the docking station. An accompanying algorithm was developed to combine the alignment subsystem, charging subsystem and the navigation subsystem of MAVEN. The algorithm consists of 3 parts: monitoring of battery levels, alignment system, and the communication with the navigational algorithm. Battery levels are monitored through the M4ATX DC-DC power supply unit. During operation, a low battery level (22V) will trigger a signal to be passed to the navigation module to bring MAVEN back to the docking station. The alignment system makes use of the holonomic capabilities of MAVEN for docking, utilising the ARToolKit which uses the marker attached onto the docking station as a guide of alignment. Charge completion is detected using a current meter installed into MAVEN, which monitors the current ﬂowing into the battery. The whole process of the docking procedure is synchronised with the navigation module to allow for a seamless and autonomous operation The prototype was tested upon with the navigation module and results suggests that the objectives of the project were met. The docking module, when paired with the navigation module, allowed for a 100% success rate in autonomous docking, charging, and subsequently undocking when a museum tour is requested. Additional utilities were added to improve the versatility of the docking station, including a lighting system to illuminate the marker in dark environments. In a separate experiment, the operational duration of the battery was obtained at about 6 hours for loaded motion, while the time taken to fully charge a drained battery amounted to about 3 hours. Recommendations for future work on the design were made to allow improvements to the current design, in order to increase the angular error of entry, and also to improve on the time taken to dock through modiﬁcation of the docking algorithm||URI:||http://hdl.handle.net/10356/64608||Rights:||Nanyang Technological University||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||MAE Student Reports (FYP/IA/PA/PI)|
checked on Sep 24, 2020
checked on Sep 24, 2020
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.