Study of Human Steering Tasks using a Neuromuscular Driver Model Naser Mehrabi Mohammad Sharif...
 Table of Contents Introduction Dynamic Modeling & Controller Design Driver Model Results and Discussion Summary and...
 Introduction  Vehicle Modeling  Driver Modeling  Simulatio...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Driver Modeling Approaches1. Bla...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Research Goals1) Develop a symbolic ...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Full Vehicle Model in MapleSim for High-...
 Introduction  Vehicle Modeling  Driver Modeling  Simulati...
 Introduction  Vehicle Modeling  Driver Modeling  ...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Driver Modeling Introducti...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Path-Following ControllerMulti-poin...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Neuromuscular System Hill Muscle...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Neuromuscular Sys...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Neuromuscular System Muscul...
 Introduction  Vehicle Modeling  Driver Modeling  Simulations Example1: Sinusoidal Steeri...
 Introduction  Vehicle Modeling  Driver Modeling ...
 Introduction  Vehicle Modeling  Driver Modeling ...
 Summary and Future Work Summary: • A high-fidelity vehicle model including a column-type electric power steer...
Questions ?18
of 18

Naser AVEC-Sep10-2012

Study of Human Steering Tasks using a Neuromuscular Driver Model
Published on: Mar 3, 2016
Published in: Automotive      
Source: www.slideshare.net


Transcripts - Naser AVEC-Sep10-2012

  • 1. Study of Human Steering Tasks using a Neuromuscular Driver Model Naser Mehrabi Mohammad Sharif Shourijeh John McPhee University of Waterloo Department of Systems Design Engineering1
  • 2.  Table of Contents Introduction Dynamic Modeling & Controller Design Driver Model Results and Discussion Summary and Future Work2
  • 3.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Electric Power Steering SystemIn a steering task, the following four systems interact: 1. Driver Resistance Torque 2. Steering system Driver Torque Driver • Mechanical system • EPS controller 3. Vehicle Steering Mechanical System 4. Environment Steer Angle Disturbance Controller Output Measurements Vehicle Dynamics EPS controller3
  • 4.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Driver Modeling Approaches1. Black box (controls) driver models  Non-Predictive  Predictive2. Neuromuscular driver models  Simple lag-delay  Pick et al.(2006), The muscles involved in the steering task are identified.   A torque-driven driver model is developed.  Hoult et al.(2008),  Activation dynamics and metabolic energy consumption are considered in a neuromuscular driver model, but EPS and vehicle dynamics has not been considered.4
  • 5.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Research Goals1) Develop a symbolic vehicle model including a column assist-type Electric Power Steering (EPS) system2) Develop a neuromuscular driver model to study interaction between a driver’s hand and steering wheel3) Design an optimal EPS controller for general population to improve steering feel4) Utilize the information from the driver model to tune a subject- specific EPS controller5
  • 6.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Full Vehicle Model in MapleSim for High-fidelity SimulationsFull vehicle model subsystems: Rear Trailing arm Suspension Front Double Wishbone Suspension Column assist-type EPS system Fiala Tire Model6
  • 7.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Schematic View of EPS System on Tfricti ular c Ang cityIncludes following components: e ri n g Wh ee l Velo Ste bc m Electric Motor  One degree freedom steering Torque Sensor (Kc) mechanism bm Km / r1 G r2  Torque sensor to measure driver torque Gea r r2 r1 Box  DC electric motor connected to the steering shaft with a gear  Rack and pinion  Viscous damping and Coulomb friction in the steering wheel and Connected to electric motor shaft vehicle Rack and Pinion7
  • 8.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Electric Power Steering Controller Design Control ObjectivesControl objectives can be Vehicle Self Alignment Torque Steering Wheel Anglecategorized as: Disturbance Dynamics Assistance  Sufficient assistant torque to the Driver Torque Steering driver. + System Steering Torque + Assist Torque Return-to-center control mode Desired  Returnability of steering wheel Current after it is released. Electric The assist PID Motor - characteristic Damping control mode Measured Current  Improve the stability of straight- line motion.8
  • 9.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Driver Modeling IntroductionDriver model can be divided into twosub-systems:  Path-following controller  Neuromuscular system Driver ModelDesired Path Actual Path Path Following Neuromuscular Vehicle Controller System Dynamics Desired Steering Angle Actual Steering Angle 9
  • 10.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Path-Following ControllerMulti-point predictive driver model [1] :  Based on a circular steady-state motion of a linear bicycle vehicle model  Considers multi-preview-point to improve the path-following properties10 [1] - Kiumars Jalali. Development of a Path-following and a Speed Control Driver Model for an Electric Vehicle, SAE Technical paper, Paper Number: 2012-01-0250, 2012.
  • 11.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Neuromuscular System Hill Muscle Model A muscle can be modeled using the following three elements: • Contractile Element (CE) = • Parallel Element (PE) CE • Series Element (SE) SE PE The force generated by the CE can be approximated by the following curves when activation (ai) is assumed unity.11
  • 12.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Neuromuscular System Hill muscle model ce f 0.5 1 1.5 ce f 1.5 1 0 v ce -10 0 10 Shortening Lengthening12
  • 13.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Neuromuscular System Musculoskeletal Driver Model Assumptions and Features  One degree of freedom system Fle (Elbow joint is assumed fixed) orx  An agonist and antagonist pair Extensor muscle at the shoulder  Constant moment arm for each muscle Fixed Joint Revolute Joint  The contractile elements have a dominant effect on the steering task Muscle 1  Activation dynamics in the CE1 neuromuscular system Vehicle Inertia Model  Indeterminate system (two muscles, CE2 one degree of freedom) θ t , θ t 13  Muscle 2
  • 14.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Example1: Sinusoidal Steering Input Inverse Dynamics IK Specified Shoulder Activations SW Kinematics Motion14
  • 15.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Closed-Loop Simulation Forward Dynamic optimization  Dynamic optimization  Optimization over the whole range of simulation  Can be used for time-dependant objectives and unknowns  Static optimization  Optimization at each instance to find the unknown activation signals  Quicker optimization time Actual Path Vehicle Steering Wheel Self Alignment Disturbance Dynamics Torque AngleDesired Path Steering System - MPPC1 NMS2 + + Steering Torque Driver + Desired Torque Assist Torque Steering Angle Desired Current Activation Signals Electric The assist PID Motor - characteristic 1 Optimization15 2 MPPC: Multi-point Predictive Controller Measured NMS: Neuromuscular System Current
  • 16.  Introduction  Vehicle Modeling  Driver Modeling  Simulations Closed-Loop Simulation Example2: ISO Double Lane Change • Vehicle velocity : 20 m/s • With MapleSim model of vehicle, EPS, and driver • Dynamic optimization to resolve muscle activations Model predictive driver model (with and without EPS) for Driver torque with and without EPS ISO double lane change path 4 4 Desired Path Without EPS 3.5 Model Predictive Driver Model 3 With EPS 3 2 Lateral Displacement (m) Driver Torque (N.m) 2.5 1 2 0 -1 1.5 -2 1 -3 0.5 -4 0 -516 -0.5 0 5 10 15 20 0 5 10 15 20 Time (s) Time(s)
  • 17.  Summary and Future Work Summary: • A high-fidelity vehicle model including a column-type electric power steering system is developed. • A two-level controller for the neuromuscular driver model is developed. • An environment for simulating driver and vehicle interactions is developed. • A method to find muscle activation signals in inverse and forward dynamics is developed. Future Work: • Incorporate the driver’s sense of steering torque into the driver model • Include metabolic energy consumption into optimization cost function • Include muscle fatigue reduction into the EPS controller’s objective • Study age, gender, and physical ability on driver’s performance17
  • 18. Questions ?18