摘要: |
This project will demonstrate the feasibility of TrainMate, a cost-effective, train station assistant robot that will primarily help disabled passengers using mobility devices to independently board and deboard trains at non-accessible, street level train stations. The project will be carried out in two stages. Work in Stage 1 work will focus on a detailed analysis of the technical requirements for the proposed system, followed by a detailed design of the electromechanical components. The overall system design will be developed along with the simulated prototype platform (base and wheelchair lift modules) and sensor network to pass all technical verifications during the software simulation in SolidWorks as per input specifications and desired use cases. This activity will involve: (1) an analysis of the requirements; (2) detailed mechanical design (prototype mechanical feasibility and design); (3) camera(s) and sensor network; (4) module integration design; and (4) technical verification and testing. Work in Stage 2 will focus on developing and validating several key algorithms (e.g., autonomous navigation, machine vision and robotic control system) which are mostly Artificial Intelligence driven. The machine vision and artificial intelligence algorithms that will be designed in Stage 2 will demonstrate a good level of accuracy for object detection after simple preliminary proof of concept and also a real-time response in simulated scenarios using sample pictures. The integration frameworks will be validated using at least 5 sample real-life scenarios. There is always a tradeoff between accuracy and speeds of the algorithms which are both essential for navigation in public transit environments. This issue will be addressed by optimizing the hardware and software algorithms according to the user’s requirements to ensure maximum possible practicality is achieved for the user desired scenarios. Finally, a fully integrated system will be developed, validated and simulated to demonstrate autonomous boarding and deboarding of a passenger using a mobility device in CAD simulated scenarios with improved efficiency, practicality and speed. The final integration of all components might pose integration related problems that could slow down the conclusion of final procedures due to extra redesign and debugging requirements. Formal documentation of test cases, pass and fail scenarios, and actions taken will help mitigate this risk. The final report will include all relevant data, methods, models, and conclusions.
The project is expected to provide the following deliverables: (1) Simulated prototype platform (base and wheelchair lift modules) and sensor network, passing all technical verifications during software simulation in SolidWorks as per input specifications and desired use cases; (2) Machine vision and artificial intelligence algorithms that can demonstrate a good level of accuracy in object detection after a preliminary proof of concept and demonstrate real-time response in simulated scenarios using sample pictures; and (3) CAD simulated scenarios with improved efficiency, practicality and speed of the fully integrated system that will demonstrate autonomous boarding and deboarding of a passenger using a mobility device. |