PROJECT ON  Intelligent Speed Adaptation Using Neuro-Fuzzy Controller GUIDED BY PROF. S S WANKHEDE Kshetrimayum Jevel Singh Lucky Amesar Nisha Kanoo Praful Kambe G.H.RAISONI COLLEGE OF ENGINEERING Department of  Electronics  & Telecommunication
OBJECTIVES This Project aims towards design of Hybrid Controller using Neuro- Fuzzy technique for Longitudinal controlling of Automotive System
Era of Automation Automation in vehicle Introduction
Why do we need controller in vehicles? WHO report on road traffic injury prevention states that 1.2 million people are killed in road accident every year 50 millions are injured Figure will increase upto 65% in next 20 years Global cost of road crashes and injuries is about US $ 518 billions per year. But the basic reason for all this is …….. “ Human error”
WHO report figure on Road accidents …..
PROBLEM ANALYSIS Driving as a continuous decision making process   Earlier speed adaptation systems with limitations Modified System HUMAN APPROACH Efficient in  Decision making NEURAL NETWORKS FUZZY LOGIC
HYBRID CONTROLLER Neural Network + Fuzzy Logic Good for learning. Not good for human to interpret its internal representation. Supervised leaning Unsupervised learning Reinforcement learning Human reasoning scheme. Fuzzy rules and membership functions are subjective. Readable Fuzzy rules Interpretable
HYBRID CONTROLLER Cont…. Neural Network + Fuzzy Logic Good for learning. Not good for human to interpret its internal representation. Supervised leaning Unsupervised learning Reinforcement learning Human reasoning scheme. Fuzzy rules and membership functions are subjective. Readable Fuzzy rules Interpretable A  Neuro-fuzzy system  is a  fuzzy system  that uses a  learning algorithm  derived from or inspired by  neural network  theory to determine its  parameters  by processing data samples.
Vehicle Controller based on  “ Generic self organizing Fuzzy-Neural Network” Mapped by   “ Yager’s  Inference Scheme”  VEHICLE  CONTROLLER It is a fuzzy-Neural network which uses Yager’s inference scheme to interpret fuzzy relations.
Yager’s Inference Scheme It is an extension of modus pones rule which is nothing but similar to implication & is also called as affirmative mode It can be stated as  If A is true  and    A  B then B is true.
STRUCTURE OF GenSOFNN Layer 1 input linguistic nodes Layer 2 input term nodes Layer 3 rule nodes Layer 4 Output term node Layer 5 output linguistic nodes
Fuzzifier Inference Engine Defuzzifier STRUCTURE OF GenSOFNN contd.. Layer 1 input linguistic nodes Layer 2 input term nodes Layer 3 rule nodes Layer 4 Output term node Layer 5 output linguistic nodes
antecedent consquent STRUCTURE OF GenSOFNN contd… DIC Technique` is used Layer 1 input linguistic nodes Layer 2 input term nodes Layer 3 rule nodes Layer 4 Output term node Layer 5 output linguistic nodes
Adaptation to vehicle controller GenSoYager GenSoYager anticipation speed Throttle speed anticipation Brake A.)  Implementation of  Longitudinal Control
Adaptation to vehicle controller B.) Training of GenSoYager system Human Driver Driving Simulator Log File Visual Feedback Action Action
INPUTS TO  VEHICLE CONTROLLER INPUTS Speed Anticipation It’s a linear variable. It depends upon speed limit of vehicle 1. It depends upon the curve & distance from curve. 2. Calculated by using log file.
Calculation of anticipation factor speed distance Anticipation algorithm anticipation
IMPLEMENTATION Problem analysis HUMAN approach Decision making GENSOFNN Training using Error back propagation algorithm & log file Provided with Fuzzy set of rules to interpret
ADVANTAGES Comparatively better control. Anticipation Factor doesn’t vary . Chances of Road mishaps reduces.
DRAWBACKS TORCS,an open source simulator which is selected for the simulation. It doesn’t take into account the action of centripetal force during the car Slipping over a turning. For this we have to depend upon the reliability of the system to control the vehicle
Scope of Work Like longitudinal control lateral control can also be implemented by using the concept of anticipation.
References M. Peden, R. Scurfield, D. Sleet, et al. World Report on road traffic injury prvention. World Health Organisation, 2004 M. Pasquier, C. Quek, and M. Toh. Fuzzylot: A Novel self-organising Fuzzy-Neural rule-based pilot system for automated vehicles.  Neural networks, vol. 14, no. 8, pp. 1099-1112, Oct. 2001. W.L.Tung, and C.Quek. GenSoFNN: A Generic self-organising Fuzzy-Neural Network . IEEE Transactions on Neural Networks, vol. 13,  no.5, pp.1075-1086, 2002
THANKING YOU QUESTIONS?

Speed adaptation using Neuro fuzzy approach

  • 1.
    PROJECT ON Intelligent Speed Adaptation Using Neuro-Fuzzy Controller GUIDED BY PROF. S S WANKHEDE Kshetrimayum Jevel Singh Lucky Amesar Nisha Kanoo Praful Kambe G.H.RAISONI COLLEGE OF ENGINEERING Department of Electronics & Telecommunication
  • 2.
    OBJECTIVES This Projectaims towards design of Hybrid Controller using Neuro- Fuzzy technique for Longitudinal controlling of Automotive System
  • 3.
    Era of AutomationAutomation in vehicle Introduction
  • 4.
    Why do weneed controller in vehicles? WHO report on road traffic injury prevention states that 1.2 million people are killed in road accident every year 50 millions are injured Figure will increase upto 65% in next 20 years Global cost of road crashes and injuries is about US $ 518 billions per year. But the basic reason for all this is …….. “ Human error”
  • 5.
    WHO report figureon Road accidents …..
  • 6.
    PROBLEM ANALYSIS Drivingas a continuous decision making process Earlier speed adaptation systems with limitations Modified System HUMAN APPROACH Efficient in Decision making NEURAL NETWORKS FUZZY LOGIC
  • 7.
    HYBRID CONTROLLER NeuralNetwork + Fuzzy Logic Good for learning. Not good for human to interpret its internal representation. Supervised leaning Unsupervised learning Reinforcement learning Human reasoning scheme. Fuzzy rules and membership functions are subjective. Readable Fuzzy rules Interpretable
  • 8.
    HYBRID CONTROLLER Cont….Neural Network + Fuzzy Logic Good for learning. Not good for human to interpret its internal representation. Supervised leaning Unsupervised learning Reinforcement learning Human reasoning scheme. Fuzzy rules and membership functions are subjective. Readable Fuzzy rules Interpretable A Neuro-fuzzy system is a fuzzy system that uses a learning algorithm derived from or inspired by neural network theory to determine its parameters by processing data samples.
  • 9.
    Vehicle Controller basedon “ Generic self organizing Fuzzy-Neural Network” Mapped by “ Yager’s Inference Scheme” VEHICLE CONTROLLER It is a fuzzy-Neural network which uses Yager’s inference scheme to interpret fuzzy relations.
  • 10.
    Yager’s Inference SchemeIt is an extension of modus pones rule which is nothing but similar to implication & is also called as affirmative mode It can be stated as If A is true and A B then B is true.
  • 11.
    STRUCTURE OF GenSOFNNLayer 1 input linguistic nodes Layer 2 input term nodes Layer 3 rule nodes Layer 4 Output term node Layer 5 output linguistic nodes
  • 12.
    Fuzzifier Inference EngineDefuzzifier STRUCTURE OF GenSOFNN contd.. Layer 1 input linguistic nodes Layer 2 input term nodes Layer 3 rule nodes Layer 4 Output term node Layer 5 output linguistic nodes
  • 13.
    antecedent consquent STRUCTUREOF GenSOFNN contd… DIC Technique` is used Layer 1 input linguistic nodes Layer 2 input term nodes Layer 3 rule nodes Layer 4 Output term node Layer 5 output linguistic nodes
  • 14.
    Adaptation to vehiclecontroller GenSoYager GenSoYager anticipation speed Throttle speed anticipation Brake A.) Implementation of Longitudinal Control
  • 15.
    Adaptation to vehiclecontroller B.) Training of GenSoYager system Human Driver Driving Simulator Log File Visual Feedback Action Action
  • 16.
    INPUTS TO VEHICLE CONTROLLER INPUTS Speed Anticipation It’s a linear variable. It depends upon speed limit of vehicle 1. It depends upon the curve & distance from curve. 2. Calculated by using log file.
  • 17.
    Calculation of anticipationfactor speed distance Anticipation algorithm anticipation
  • 18.
    IMPLEMENTATION Problem analysisHUMAN approach Decision making GENSOFNN Training using Error back propagation algorithm & log file Provided with Fuzzy set of rules to interpret
  • 19.
    ADVANTAGES Comparatively bettercontrol. Anticipation Factor doesn’t vary . Chances of Road mishaps reduces.
  • 20.
    DRAWBACKS TORCS,an opensource simulator which is selected for the simulation. It doesn’t take into account the action of centripetal force during the car Slipping over a turning. For this we have to depend upon the reliability of the system to control the vehicle
  • 21.
    Scope of WorkLike longitudinal control lateral control can also be implemented by using the concept of anticipation.
  • 22.
    References M. Peden,R. Scurfield, D. Sleet, et al. World Report on road traffic injury prvention. World Health Organisation, 2004 M. Pasquier, C. Quek, and M. Toh. Fuzzylot: A Novel self-organising Fuzzy-Neural rule-based pilot system for automated vehicles. Neural networks, vol. 14, no. 8, pp. 1099-1112, Oct. 2001. W.L.Tung, and C.Quek. GenSoFNN: A Generic self-organising Fuzzy-Neural Network . IEEE Transactions on Neural Networks, vol. 13, no.5, pp.1075-1086, 2002
  • 23.