High Fidelity Modelling of Motorcycle Dynamics

Modern racing and high performance road motorcycles have begun to use ever more sophisticated control systems to improve performance and safety. Traction control, anti-wheelie control, engine brake control and in the case of road bikes, anti-lock braking are now ubiquitous. Electronically adjustable suspension damping is becoming ever more common. At the time of writing, no manufacturer uses a high fidelity real time model of the motorcycle and tyres in their control strategies. The use of such a model would assist in integrating separate strategies into an overall vehicle stability strategy. Furthermore, the development of racing and high performance road motorcycles has traditionally depended more on experience and testing than on modelling and simulation. One reason for this has been the lack of availability of a high fidelity motorcycle model suitable for the task. The development of general multibody simulation software such as MSC Adams has been driven primarily by the four wheel world and is much more suited to cars than motorcycles. Simplified or over-linearised models lack the required fidelity. A new high fidelity multibody motorcycle model for use in real time is proposed. It is an eleven degree of freedom model and the equations of motion are formulated using Kane's equations. Kane's method has been chosen over other methodologies as it scales well to complex systems and lends itself to automated numerical computation in a way that other methods do not. There are few linear relationships in the mechanism of a motorcycle, but many relationships are very nearly linear. This fact is used to simplify the model while maintaining fidelity and this contributes to lowering the computation cost. The tyre model is the widely used Magic Formula model, developed by Pacejka and others. The motorcycle model estimates data which is not directly measurable, such as tyre shear forces and slip angles. A motorcycle was instrumented and data was recorded during various tests, including coast-down testing and some very competitive laps of a race circuit by a professional motorcycle racer and test rider. The data from the coast-down tests was used to estimate aerodynamic and engine drag parameters. The model was validated against the race track test data with satisfactory results. Tyre model parameters were not available for the tyres used in testing, so a method was developed to estimate the tyre parameters which predict lateral force. The behaviour of the model was analysed by applying various test functions and root locus analysis was carried out to investigate the system modes. The utility of the model for motorcycle development and set-up was investigated and a method was developed whereby motorcycle parameters may be optimised with respect to certain performance criteria. This method may be used to explore directions for future development and also to optimise settings for racing