Role of vehicle dynamic modeling fidelity with haptic collaboration in steer by wire systems
Naik, Anand P.
MetadataShow full item record
Steer-By-Wire (SBW) systems offer many benefits ranging from mechanical isolation of steering wheel from the road, weight reduction in the steering system, relaxed packaging constraints, to facilitating advanced driving-safety systems. However, there is also the loss of proprioception ("road feel") which is a critical feedback element for manual vehicle control. To this end, haptic interfaces for SBW systems have been proposed to restore the intimacy of interactive control back to the driver. Candidate solutions for mimicking the steering feel have ranged from direct instrumented-pickup and feedback of road-wheel interactions (using accelerometers/force-sensors) to steering torque prediction schemes based on mathematical dynamics models (of tire-road, suspension, power-steering systems) in conjunction with selected real-time measurements. While the latter approach offers the most promise, real-time implementations at high sampling rates in noisy environments pose challenges. A careful selection of fidelity of the underlying dynamic model as well as good matching of haptic model-device capabilities is critical. The degree of realism for the user-vehicle interaction is dependent on the fidelity of the underlying computational vehicle dynamics model. Hence, in this work we focus on creation, implementation and preliminary testing of varying fidelity vehicle-dynamic models for haptic steer-by-wire driving tasks. Additionally the SBW paradigm can simplify implementation of shared/collaborative control (steering) of the underlying mechanical system (vehicle). In this thesis we implement and evaluate various possibilities for sharing of control between multiple individual users and/or between user and automation technology. Quantitative performance evaluation is conducted to understand the role of vehicle dynamics modeling fidelity for haptic SBW tasks along with the evaluation of 3 modes of shared control, user automation control vs. individual control. In particular, preliminary experimental analyses with five subjects using three performance metrics (Error Value Parameter, FFT Power Ratio and Free Control Oscillations) were evaluated to quantify vehicle models and collaboration modes performance.