In addition to the engineering knowledge base that has been traditionally taught, today's undergraduate engineering students need to be given the opportunity to practice a set of skills that will be demanded to them by future employers, namely: creativity, teamwork, problem solving, leadership and the ability to generate innovative ideas. In order to achieve this and educate engineers with both in-depth technical knowledge and professional skills, universities must carry out their own innovating and find suitable approaches that serve their students. This article presents a novel approach that involves university-industry collaboration. It is based on creating a student community for a particular company, allowing students to deal with real industry projects and apply what they are learning in the classroom. A sample project for the German sports brand adidas is presented, along with the project results and evaluation by students and teachers. The university-industry collaborative approach is shown to be beneficial for both students and industry.

The accuracy of multi-body simulation results relies on the model building process and the accuracy of the model parameters. A more widespread use of vehicle dynamic calculations in the acceptance process would be possible if the validation of the model was secured. This paper proposes a methodology for an objective and direct identification of the values of the parameters of a rail vehicle model, using the results of the stationary tests defined in the acceptance process of railway vehicles (EN14363). The methodology also takes into account the variability of the measuring process by providing a probabilistic estimation of the identified parameters. The methodology is validated using an example of a virtual wheel unloading test (simulation). Four significant model parameters can be accurately calculated: vertical primary and secondary suspension stiffness, stiffness of the anti-roll bar, and height of the null moment point (the lateral/roll coupling effect of the air spring). Finally, a reduction method is shown which decreases the uncertainties of the identified parameters by up to 50%.