Optimal model predictive control of a commuter train system

通勤列车广泛应用于地球仪的城市公共交通系统。这些列车中的许多都配备了自动速度控制器，使它们能够在很少或没有操作员干预的情况下自动在目的地之间移动。然而，这些控制算法通常被设计成使得它们可以在具有有限计算资源的常规车载列车计算机上运行。这反过来又大大限制了它们的性能和适应不断变化的操作条件的能力。该项目汇集了来自麦克马斯特大学和泰雷兹加拿大运输解决方案（TCTS）的研究人员和工程师团队，旨在为电动通勤列车开发先进的自动控制器，以解决现有控制器的一些缺点。新的控制器将基于最优模型预测控制理论，预计将在性能和对系统不确定性的鲁棒性方面产生显着改善。能耗、乘客舒适度和乘坐质量以及运营成本是这些新控制器将优化的一些性能指标。控制器还将能够真实的适应列车和轨道特性的任何变化，以保持其高性能并避免潜在的系统不稳定。这一功能还有望大大减少TCTS工程师在现场调整列车速度控制系统的时间。研究团队还将研究控制算法和基于图形处理器单元的并行计算的有效实施，以实现新列车控制器计算的实时执行。 加拿大将受益于这一项目所产生的研究与发展和商业化活动，并受益于在对广大加拿大公司具有重要意义的领域培训一些高素质人员。此外，加拿大的运输系统可能利用这项研究所产生的技术，预计会产生重大的经济、环境和社会效益。

Commuter trains are widely used in urbane mass transit systems around the globe. Many of these trains are equipped with automatic speed controllers that enable them to autonomously move between destinations with little or no operator intervention. However, these control algorithms have often been designed so they can run on conventional on-board train computers with limited computational resources. This, in turn, has substantially restricted their performance and the their ability to adapt to changing operating conditions. This project brings together a team of researchers and engineers from McMaster University and Thales Canada Transportation Solutions (TCTS) with the aim of developing advanced automatic controllers for electric commuter trains to address some of the shortcomings of existing controllers. The new controllers will be based on the theory of optimal model predictive control and are expected to yield significant improvements in terms of performance and robustness to system uncertainty. Energy consumption, passenger comfort and ride quality, and the cost of operation are some of the performance metrics that will be optimized by these new controllers. The controllers will also be able to adapt, in real time, to any changes in the train and track characteristics to maintain their high performance and avoid potential system instability. This feature is also expected to significantly reduce the time the TCTS engineers would spend tuning the train speed control system in the field. The research team will also investigate efficient implementation of the control algorithms and Graphic Processor Unit-based parallel computing to enable real-time execution of the new train controllers computations. Canada will benefit from the research and development, and commercialization activities arising from this project, as well as from the training of a number of highly qualified personnel in areas of importance to a broad spectrum of Canadian companies. In addition, significant economic, environmental and social benefits are expected from the potential use of the resulting technologies from this research in Canada's transit systems.