Homework: optimal control without a reference trajectory A disadvantage of the formulation that we developed in class is that it requires a reference trajectory.

Slides:



Advertisements
Similar presentations
Visual feedback in the control of reaching movements David Knill and Jeff Saunders.
Advertisements

Learning Theory Reza Shadmehr Optimal feedback control stochastic feedback control with signal dependent noise.
Numerical Solution of Linear Equations
Chapter 7: Matrix Algebra 1.(7.1) Matrix Arithmetic a)Matrix-Vector Multiplication b)Matrix-Matrix Multiplication 2.(7.2) Applications 1.(7.1) Matrix Arithmetic.
Partial Differential Equations
Gaussian Elimination Matrices Solutions By Dr. Julia Arnold.
Linear Algebra Applications in Matlab ME 303. Special Characters and Matlab Functions.
Chapter 8: Long-Term Dynamics or Equilibrium 1.(8.1) Equilibrium 2.(8.2) Eigenvectors 3.(8.3) Stability 1.(8.1) Equilibrium 2.(8.2) Eigenvectors 3.(8.3)
1.2 Row Reduction and Echelon Forms
Lecture 11: Recursive Parameter Estimation
Quantifying Generalization from Trial-by-Trial Behavior in Reaching Movement Dan Liu Natural Computation Group Cognitive Science Department, UCSD March,
MAE 552 – Heuristic Optimization Lecture 8 February 8, 2002.
ECIV 301 Programming & Graphics Numerical Methods for Engineers Lecture 2 Mathematical Modeling and Engineering Problem Solving.
CS274 Spring 01 Lecture 5 Copyright © Mark Meyer Lecture V Higher Level Motion Control CS274: Computer Animation and Simulation.
KINEMATICS ANALYSIS OF ROBOTS (Part 1) ENG4406 ROBOTICS AND MACHINE VISION PART 2 LECTURE 8.
CIS 540 Principles of Embedded Computation Spring Instructor: Rajeev Alur
Lecture 10 CSS314 Parallel Computing
A kinematic cost Reza Shadmehr. Subject’s performanceMinimum jerk motion Flash and Hogan, J Neurosci 1985 Point to point movements generally exhibit similar.
Colorado Center for Astrodynamics Research The University of Colorado 1 STATISTICAL ORBIT DETERMINATION Satellite Tracking Example of SNC and DMC ASEN.
We will use Gauss-Jordan elimination to determine the solution set of this linear system.
Matlab Basics Tutorial. Vectors Let's start off by creating something simple, like a vector. Enter each element of the vector (separated by a space) between.
Department of Electrical Engineering, Southern Taiwan University Robotic Interaction Learning Lab 1 The optimization of the application of fuzzy ant colony.
Vectors and Matrices In MATLAB a vector can be defined as row vector or as a column vector. A vector of length n can be visualized as matrix of size 1xn.
Water Resources Development and Management Optimization (Linear Programming) CVEN 5393 Mar 4, 2011.
OR Perturbation Method (tableau form) (after two iterations, optimal solution obtained) (0+2  1 ) (0+2  2 ) (1+  3 )
Optimal Therapy After Stroke: Insights from a Computational Model Cheol Han June 12, 2007.
Chapter 4 Review: Manipulating Matrices Introduction to MATLAB 7 Engineering 161.
AIM: How do we perform basic matrix operations? DO NOW:  Describe the steps for solving a system of Inequalities  How do you know which region is shaded?
Stefanos Zafeiriou Machine Learning(395) Course 395: Machine Learning – Math. Intro. Brief Intro to Matrices, Vectors and Derivatives: Equality: Two matrices.
Human-Computer Interaction Kalman Filter Hanyang University Jong-Il Park.
Learning to perceive how hand-written digits were drawn Geoffrey Hinton Canadian Institute for Advanced Research and University of Toronto.
Day 3 Markov Chains For some interesting demonstrations of this topic visit: 2005/Tools/index.htm.
Solution of a System of ODEs with POLYMATH and MATLAB, Boundary Value Iterations with MATLAB For a system of n simultaneous first-order ODEs: where x is.
Statistical learning and optimal control: A framework for biological learning and motor control Lecture 4: Stochastic optimal control Reza Shadmehr Johns.
Learning Theory Reza Shadmehr Optimal feedback control stochastic feedback control with and without additive noise.
A Neural Model for the Adaptive Control of Saccadic Eye Movements Sohrab Saeb, Cornelius Weber and Jochen Triesch International Joint Conference on Neural.
Class 7: Answers 1 (C) Which of the following matrices below is in reduced row echelon form? A B C D. None of them.
CONSTANT EFFORT COMPUTATION AS A DETERMINANT OF MOTOR BEHAVIOR Emmanuel Guigon, Pierre Baraduc, Michel Desmurget INSERM U483, UPMC, Paris, France INSERM.
ECE 450 Introduction to Robotics Section: Instructor: Linda A. Gee 10/07/99 Lecture 11.
ROBOTIC ARM 2 Wilmer Arellano © Hardware  Next slide shows sensor connection to analog pin 0 and Motor 1 connection. Lecture is licensed under.
Trajectory Generation
Smart Icing System Review, September 30 – October 1, 2002 Autopilot Analysis and EP Scheme for the Twin Otter under Iced Conditions. Vikrant Sharma University.
FORECASTING METHODS OF NON- STATIONARY STOCHASTIC PROCESSES THAT USE EXTERNAL CRITERIA Igor V. Kononenko, Anton N. Repin National Technical University.
Graphics Lecture 2: Slide 1 Lecture 2 Transformations for animation.
University of Colorado Boulder ASEN 5070 Statistical Orbit determination I Fall 2012 Professor George H. Born Professor Jeffrey S. Parker Lecture 9: Least.
Lab 9: practice with functions Some tips to make your functions a little more interesting.
© 2016 Pearson Education, Ltd. Linear Equations in Linear Algebra LINEAR MODELS IN BUSINESS, SCIENCE, AND ENGINEERING.
Department of Electrical Engineering, Southern Taiwan University 1 Robotic Interaction Learning Lab The ant colony algorithm In short, domain is defined.
AAE 556 Aeroelasticity Lecture 7 – Control effectiveness (2)
CIS 540 Principles of Embedded Computation Spring Instructor: Rajeev Alur
Matrices Rules & Operations.
Newton’s second law In this lesson, students learn to apply Newton's second law to calculate forces from motion, and motion from forces. The lesson includes.
Single-Joint Movements
Newton’s second law In this lesson, students learn to apply Newton's second law to calculate forces from motion, and motion from forces. The lesson includes.
Degrees Radians radians = degrees degrees = radians.
Newton’s second law Pg. 21 in NB
Water Resources Development and Management Optimization (Linear Programming) CVEN 5393 Mar 7, 2011.
Advanced Algorithms Analysis and Design
Perturbation method, lexicographic method
Motor Control Theories
Albert C. Chaney 24 January 2008 Dynamics and Control Initial Controller Design AAE 450 Spring 2008.
Homework 2 This homework is the first use of quantum gates. In lectures we learned about the following gates: inverter, Feynman (controlled NOT), Toffoli.
دانشگاه صنعتي اميركبير
Consider Covariance Analysis Example 6.9, Spring-Mass
Coordinate Transformation in 3D Final Project Presentation
Introduction to projectile motion MANDATORY experiment
AAE 556 Aeroelasticity Lecture 8
1st Annual Israel Multinational BMD Conference & Exhibition
Matrix Multiplication Sec. 4.2
Presentation transcript:

Homework: optimal control without a reference trajectory A disadvantage of the formulation that we developed in class is that it requires a reference trajectory r. It would be nice if we could get rid of this vector in our cost function. Suppose that we want to move a limb to targets that can appear anywhere. We want an optimal control law that can generate the motor commands no matter where the goal of the movement might be. To approach the problem, let us extend the state vector to include information about where the goal is. Below, g represents location of the goal. We want an optimal control law of this form. Where: Because we got rid of r, we now have:

Using the model of the elbow described in the last lecture, here we intend to simulate control of the arm when the target changes midway into the movement. Start with x=[pos, velocity] and the parameter values for k, b, and m and compute matrix A and C for delta_t of 0.01 sec. Now let us extend x so it is [pos, velocity, goalpos]. Add a row to A and C so that goal position remains invariant in time and motor commands do not affect the goal position. Using B from the last slide, we have a y that is a 2x1. Suppose we want a movement time of 0.45 sec with the following cost characteristics: Solve for the sequence of W matrices. You now have a control law.

1.Suppose we see a target at 30 deg (you will need to do your simulation in radians). The initial conditions are: x=[ ]. Simulate the movement and plot position and the motor commands. To do the simulation, on each iteration you will need to start with a measure of state and generate a motor command. Based on the motor command, predict the next state, and continue with the loop. In this simulation, the goal remains the same. 2.Now suppose that during the movement, at 70ms into the movement the target jumps from 30 deg to 15 deg. Simulate this. Note that the system nearly completely compensates for the perturbation. 3.Let’s investigate how the compensation depends on the timing of the goal change. Suppose that the goal changes at 150ms into the movement. Simulate this. You should see that the compensation is partial. 4.Finally, change the goal at 220ms. You should see that there is very little compensation. People’s movements show similar behavior. When target of a reach changes very early in the movement, there is complete compensation. However, a late change in the movement is often under-compensated.