Exam Syllabus

Related Courses

  • CS556- Robotics
  • CS657-Intelligent Systems and Control and
  • Either CS550 Artificial Intelligence OR CS553 Neural Networks (i.e. students have a choice of answering CS550 and/or CS553 related questions. These two courses individually or in combination constitute about one third of the exam)

References

  1. M. Negnevitsky, Artificial Intelligence – A guide to intelligent systems Addison-Wesley, 2005, Chapters 2, 4, 7.
  2. M. Tarokh, Intelligent Systems and Control, Lectures
  3. J. J. Craig, Introduction to Robotics, Addison Wesley Publishers, 2005, Chapters 1,2,7,8,9.
  4. M. Vuskovic, Lecture Notes for CS556 and CS656 (only parts)
  5. S Hayking, Neural Networks, Prentice Hall, 2nd edition.
  6. Artificial Intelligence: Structures and Strategies for Complex Problem-Solving, 5th edition, Addison-Wesley, 2005.

 

Topics

    1. Intelligent Systems and Control
      1. General rule bas expert systems (structure, characteristics, chaining inferences, conflict resolution). (Ref [1], Chapter 2).
      2. Fuzzy expert systems (fuzzy sets and their operations, linguistic variable, fuzzy rules, fuzzy inference, defuzzification, applications), (Ref[2], Chapter 4).
      3. Evolutionary and genetic algorithms (simulation of natural evolution, genetic algorithms, genetic operators, fitness function, applications), (Ref[2], Chapter [7]).
      4. Simulation and control of dynamics systems (modeling using Simulink, PID controllers, implementation, fuzzy control, stability and performance evaluations), (Ref[2]).

 

    1. Robotics (Refs [3] and [4])
      1. Basic concepts of robotics (Candidates must be able to describe the terms: joints (revolute, prismatic), links, degrees of freedom, end-effectors, position and force/torque sensors, encoders, actuators, joint controllers, forward and inverse kinematics, velocity mapping, Cartesian and joints space,
      2. Forward kinematics of n DOF planar robots (Candidates should be able to write equations for position and orientation of a planar serial manipulator of any number of joints and to solve simple problems associated with the equations)
      3. Inverse kinematics of 3 DOF planar robots (Candidate should be able to derive solutions for joint angles given Cartesian position and orientation for a planar serial manipulator with three joints, including the resolution of the ambiguity of solutions.)
      4. Static forces and velocities (Jacobian) (Candidates should be able to compute the Cartesian velocities or joint torques given joint velocities or Cartesian forces and torques by using the manipulator Jacobian. In addition the candidate should be able to derive the wrist Jacobian of a planar serial manipulator.)
      5. Independent robot joint controllers (Candidates should be able to describe a simplified 2nd order linear model of a robot actuator based on DC motor with permanent magnet, and to calculate the parameters of the independent PD joint controller.
      6. Trajectory generators (Candidates should be able to discuss the purpose and types of various velocity profiles for trajectory generation, including cubic, quintic, cycloid and trapezoidal profiles, essay questions)

 

    1. Neural Networks
      1. Learning processes (Ref [5], Chapter 2)
      2. Single layer perceptron (Ref [5], Chapter 3)
      3. Multilayer perceptron (Ref [5], Chapter 4)
      4. Radial basis function neural networks (Ref [5], Chapter 4)
      5. Self organizing map neural networks Learning vector quantization neural networks (Ref [5], Chapter 9)

 

  1. Artificial Intelligence
    1. Logic as a Representation Language (Ref [6],Chapter 2, Sec. 2-6, pp. 50-78, questions 5-12)
      • first order predicate calculus
      • interpretations and truth value assignment, quantification
      • inference rules, soundness, completeness, unification
      • examples (e.g. robotics blocks world, financial advisor)
    2. Search Strategies (Ref[6], Chapter 3, Sec. 1.3-5, pp. 87-122, questions 2,4-12)
      • state space search structures (graphs, bfs, dfs, backtracking)
      • data-driven search (forward chaining)
      • goal-driven search (backward chaining)
      • and/or graphs and examples (e.g. English grammar, financial advisor)
    3. Heuristic Search (Ref [6], Chapter 4, Sec. 2,3,5,7, pp. 133-150, 157-160, 162-163, questions 4-12)
      • best-first search, the A* algorithm, beam search
      • sample heuristics, horizon effect/selective deepening
      • admissibility, monotonicity, informedness
      • tradeoff between search complexity and heuristic complexity
    4. Rule-Based Problem-Solving (Ref [6], Chapter 6, Sec. 2,5, pp.200-217, questions 5,7-10)
      • examples (e.g. 8-puzzle, knight’s tour, etc.)
      • data-driven and goal-driven use
      • advantages of this problem-solving architecture
    5. Knowledge Representation (Ref[6], Chapter 7, Sec 1.1,1.2,2, pp. 228-234, 248-258, 273-276, questions 1-10, 13-17)
      • semantic networks and network representations
      • conceptual graphs
      • generalization, specialization, propositions
      • equivalence to logic