PTW rider modelling

The literature on PTW rider modelling is not as rich as for car driver modelling. Fundamental work was proposed in the 1970s; for example, by Weir [Wei72] and Rice [Ric78]. Their main concern was to determine the extent to which steering torque and upper body inclination are involved in the control of both the vehicle trajectory and tilt angle. There has been a renewed interest in the modelling (and understanding) of the dynamics of PTW riders over the past ten years, as shown in an extensive review by Popov et al.
(2010). The main motivation of this studies remains the fundamental understanding of the rider control strategy. For example, what kind of action is prevalent with regard to steering torque, upper body inclination or the force applied on the foot pegs and the tank? Another more practical goal is the development of a virtual driver, something that has become more and more necessary with the development of PTW virtual design for use in simulator pilot training [MYA02].
As shown by Popov et al. [RPM10], two approaches co-exist, as is the case for car driver modelling. The first approach attempts to set up “black box”-type models that use well-known optimal control tools, often with a preview function. Based on the assumption that the control strategy of car drivers or PTW riders relies on the anticipation of the curvature of the road at various points, the objective is to design an optimal (often discrete-time) controller with a preview function. Such a controller is assumed to have the same “inputoutput” behaviour that applies to humans. The second approach, which is closer to ours, attempts to explicitly model the various sensory, cognitive and neuromuscular human functions. To our knowledge, however, no work has succeeded in proposing a synthetic model of the sensorimotor processes implied in PTW control that integrates the various sensory inputs used by the rider (visual, vestibular and haptic cues) and the biomechanics of the postural adjustments that determine how the rider tilts the PTW.
Recently, a similar approach to driver modelling was successfully applied by IRCCyN (Control and PsyCoTec teams) within the  framework of the ANR VTT PARTAGE project. The model explicitly
represents the visual control of steering (anticipation and  compensation) and a physiologically plausible neuromuscular system (Fig. 2). The model was designed and experimentally validated in a driving simulator [MSCCL11], [SCMLC11]. It was then used in the design of an automatic co-pilot, which shared control of the steering wheel with the driver in a continuous way [Sal12], [SCCLM12], [SCCLM13]. Research proposed within VIROLO+ aims to generalize this work to the PTW control activity. This will include, in particular, the vestibular perception of gravito-inertial forces and the dynamic management of the various forces exerted by the rider on the vehicle based on a biomechanical model of the posture.