simiam.controller.FollowWall

classdef FollowWall < simiam.controller.Controller

% Copyright (C) 2013, Georgia Tech Research Corporation
% see the LICENSE file included with this software

    properties
        
        % memory banks
        E_k
        e_k_1
        
        % gains
        Kp
        Ki
        Kd
        
        % plot support
        p
        
        % sensor geometry
        calibrated
        sensor_placement
        
        v_t
        v_p
        v_pt
        
        d_fw
    end
    
    properties (Constant)
        inputs = struct('v', 0, 'direction', 'right');
        outputs = struct('v', 0, 'w', 0)
    end
    
    methods
        
        function obj = FollowWall()
            obj = [e-mail]('follow_wall');            
            obj.calibrated = false;
            
            obj.Kp = 5;
            obj.Ki = 0.01;
            obj.Kd = 0.1;
            
            obj.E_k = 0;
            obj.e_k_1 = 0;
            
            %% START CODE BLOCK %%
            obj.d_fw = 0.15;
            %% END CODE BLOCK %%
            
%             obj.p = simiam.util.Plotter();
        end
        
        function outputs = execute(obj, robot, state_estimate, inputs, dt)
            
            % Compute the placement of the sensors
            if(~obj.calibrated)
                obj.set_sensor_geometry(robot);
            end
            
            % Unpack state estimate
            [x, y, theta] = state_estimate.unpack();
            
            % Poll the current IR sensor values 1-9
            ir_distances = robot.get_ir_distances();
                        
            % Interpret the IR sensor measurements geometrically
            ir_distances_rf = obj.apply_sensor_geometry(ir_distances, state_estimate);            
            
            % Compute the heading vector
            % 1. Select p_2 and p_1, then compute u_fw_t
            if(strcmp(inputs.direction,'right'))
                % Pick two of the right sensors based on ir_distances
                S = [1:4 ; ir_distances(8:-1:5)];
                [Y,i] = sort(S(2,:));
                S = S(1,i);
                
                Sp = 8:-1:5;
                
                S1 = Sp(S(1));
                S2 = Sp(S(2));
                
                if(S1 < S2)
                    p_1 = ir_distances_rf(:,S2);
                    p_2 = ir_distances_rf(:,S1);
                else
                    p_1 = ir_distances_rf(:,S1);
                    p_2 = ir_distances_rf(:,S2);
                end
                
            else
                % Pick two of the left sensors based on ir_distances
                S = [1:4 ; ir_distances(1:4)];
                [Y,i] = sort(S(2,:));
                S = S(1,i);
                if(S(1) > S(2))
                    p_1 = ir_distances_rf(:,S(2));
                    p_2 = ir_distances_rf(:,S(1));
                else
                    p_1 = ir_distances_rf(:,S(1));
                    p_2 = ir_distances_rf(:,S(2));
                end
            end
            
            u_fw_t = p_2-p_1;

            % 2. Compute u_a, u_p, and u_fw_tp to compute u_fw_p
            
            u_fw_tp = u_fw_t/norm(u_fw_t);
            u_a = p_1;
            u_p = [x;y];
            
            u_fw_p = ((u_a-u_p)-((u_a-u_p)'*u_fw_tp)*u_fw_tp);
            
            % 3. Combine u_fw_tp and u_fw_pp into u_fw;
            u_fw_pp = u_fw_p/norm(u_fw_p);
            u_fw = obj.d_fw*u_fw_tp+(u_fw_p-obj.d_fw*u_fw_pp);
            
            
            % Compute the heading and error for the PID controller
            theta_fw = atan2(u_fw(2),u_fw(1));
            e_k = theta_fw-theta;
            e_k = atan2(sin(e_k),cos(e_k));
                                    
            e_P = e_k;
            e_I = obj.E_k + e_k*dt;
            e_D = (e_k-obj.e_k_1)/dt;
              
            % PID control on w
            v = inputs.v;
            w = obj.Kp*e_P + obj.Ki*e_I + obj.Kd*e_D;
            
            % Save errors for next time step
            obj.E_k = e_I;
            obj.e_k_1 = e_k;
                        
            % plot
            obj.p.plot_2d_ref(dt, atan2(sin(theta),cos(theta)), theta_fw, 'c');
            
%             fprintf('(v,w) = (%0.4g,%0.4g)\n', v,w);            
            % velocity control
            v = 0.25/(log(abs(w)+2)+1);
            outputs.v = v;
            outputs.w = w;
        end
        
        % Helper functions
        
        function ir_distances_rf = apply_sensor_geometry(obj, ir_distances, state_estimate)
                    
            % Apply the transformation to robot frame.
            
            ir_distances_sf = zeros(3,9);
            for i=1:9
                x_s = obj.sensor_placement(1,i);
                y_s = obj.sensor_placement(2,i);
                theta_s = obj.sensor_placement(3,i);
                
                R = obj.get_transformation_matrix(x_s,y_s,theta_s);
                ir_distances_sf(:,i) = R*[ir_distances(i); 0; 1];
            end
            
            % Apply the transformation to world frame.
            
            [x,y,theta] = state_estimate.unpack();
            
            R = obj.get_transformation_matrix(x,y,theta);
            ir_distances_rf = R*ir_distances_sf;
            
            ir_distances_rf = ir_distances_rf(1:2,:);
        end
        
        function set_sensor_geometry(obj, robot)
            obj.sensor_placement = zeros(3,9);
            for i=1:9
                [x, y, theta] = robot.ir_array(i).location.unpack();
                obj.sensor_placement(:,i) = [x; y; theta];
            end                        
            obj.calibrated = true;
        end
        
        function R = get_transformation_matrix(obj, x, y, theta)
            R = [cos(theta) -sin(theta) x; sin(theta) cos(theta) y; 0 0 1];
        end
        
    end
    
end