| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% | 
					
						
							|  |  |  | % GTSAM Copyright 2010, Georgia Tech Research Corporation,  | 
					
						
							|  |  |  | % Atlanta, Georgia 30332-0415 | 
					
						
							|  |  |  | % All Rights Reserved | 
					
						
							|  |  |  | % Authors: Frank Dellaert, et al. (see THANKS for the full author list) | 
					
						
							|  |  |  | %  | 
					
						
							|  |  |  | % See LICENSE for the license information | 
					
						
							|  |  |  | % | 
					
						
							|  |  |  | % @brief Simple robotics example using the pre-built planar SLAM domain | 
					
						
							|  |  |  | % @author Alex Cunningham | 
					
						
							|  |  |  | % @author Frank Dellaert | 
					
						
							|  |  |  | % @author Chris Beall | 
					
						
							|  |  |  | % @author Vadim Indelman | 
					
						
							|  |  |  | % @author Can Erdogan | 
					
						
							|  |  |  | %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2012-08-06 03:31:27 +08:00
										 |  |  | import gtsam.* | 
					
						
							|  |  |  | 
 | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | %% Assumptions | 
					
						
							|  |  |  | %  - All values are axis aligned | 
					
						
							|  |  |  | %  - Robot poses are facing along the X axis (horizontal, to the right in images) | 
					
						
							|  |  |  | %  - We have full odometry for measurements | 
					
						
							|  |  |  | %  - The robot is on a grid, moving 2 meters each step | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | %% Create graph container and add factors to it | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | graph = NonlinearFactorGraph; | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | 
 | 
					
						
							|  |  |  | %% Add prior | 
					
						
							|  |  |  | % gaussian for prior | 
					
						
							| 
									
										
										
										
											2012-07-18 23:47:06 +08:00
										 |  |  | priorNoise = noiseModel.Diagonal.Sigmas([0.3; 0.3; 0.1]); | 
					
						
							|  |  |  | priorMean = Pose2(0.0, 0.0, 0.0); % prior at origin | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | graph.add(PriorFactorPose2(1, priorMean, priorNoise)); % add directly to graph | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | 
 | 
					
						
							|  |  |  | %% Add odometry | 
					
						
							|  |  |  | % general noisemodel for odometry | 
					
						
							| 
									
										
										
										
											2012-07-18 23:47:06 +08:00
										 |  |  | odometryNoise = noiseModel.Diagonal.Sigmas([0.2; 0.2; 0.1]); | 
					
						
							|  |  |  | odometry = Pose2(2.0, 0.0, 0.0); % create a measurement for both factors (the same in this case) | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | graph.add(BetweenFactorPose2(1, 2, odometry, odometryNoise)); | 
					
						
							|  |  |  | graph.add(BetweenFactorPose2(2, 3, odometry, odometryNoise)); | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | 
 | 
					
						
							|  |  |  | %% Add measurements | 
					
						
							|  |  |  | % general noisemodel for measurements | 
					
						
							| 
									
										
										
										
											2012-07-18 23:47:06 +08:00
										 |  |  | measurementNoise = noiseModel.Diagonal.Sigmas([0.1; 0.2]); | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | 
 | 
					
						
							|  |  |  | % print | 
					
						
							|  |  |  | graph.print('full graph'); | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | %% Initialize to noisy points | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | initialEstimate = Values; | 
					
						
							|  |  |  | initialEstimate.insert(1, Pose2(0.5, 0.0, 0.2)); | 
					
						
							|  |  |  | initialEstimate.insert(2, Pose2(2.3, 0.1,-0.2)); | 
					
						
							|  |  |  | initialEstimate.insert(3, Pose2(4.1, 0.1, 0.1)); | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | 
 | 
					
						
							|  |  |  | initialEstimate.print('initial estimate'); | 
					
						
							|  |  |  | 
 | 
					
						
							|  |  |  | %% set up solver, choose ordering and optimize | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | params = DoglegParams; | 
					
						
							|  |  |  | params.setAbsoluteErrorTol(1e-15); | 
					
						
							|  |  |  | params.setRelativeErrorTol(1e-15); | 
					
						
							|  |  |  | params.setVerbosity('ERROR'); | 
					
						
							|  |  |  | params.setVerbosityDL('VERBOSE'); | 
					
						
							| 
									
										
										
										
											2014-05-26 05:46:47 +08:00
										 |  |  | params.setOrdering(graph.orderingCOLAMD()); | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | optimizer = DoglegOptimizer(graph, initialEstimate, params);                       | 
					
						
							| 
									
										
										
										
											2012-02-06 11:33:40 +08:00
										 |  |  | 
 | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | result = optimizer.optimizeSafely(); | 
					
						
							| 
									
										
										
										
											2012-02-06 11:33:40 +08:00
										 |  |  | result.print('final result'); | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | 
 | 
					
						
							|  |  |  | %% Get the corresponding dense matrix | 
					
						
							| 
									
										
										
										
											2014-05-26 05:46:47 +08:00
										 |  |  | ord = graph.orderingCOLAMD(); | 
					
						
							|  |  |  | gfg = graph.linearize(result); | 
					
						
							|  |  |  | denseAb = gfg.augmentedJacobian; | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | 
 | 
					
						
							|  |  |  | %% Get sparse matrix A and RHS b | 
					
						
							|  |  |  | IJS = gfg.sparseJacobian_(); | 
					
						
							|  |  |  | Ab=sparse(IJS(1,:),IJS(2,:),IJS(3,:)); | 
					
						
							|  |  |  | A = Ab(:,1:end-1); | 
					
						
							|  |  |  | b = full(Ab(:,end)); | 
					
						
							| 
									
										
										
										
											2012-07-28 03:02:11 +08:00
										 |  |  | clf | 
					
						
							| 
									
										
										
										
											2012-01-31 06:00:13 +08:00
										 |  |  | spy(A); | 
					
						
							| 
									
										
										
										
											2012-07-24 03:21:05 +08:00
										 |  |  | title('Non-zero entries in measurement Jacobian'); |