| 1 | /* ---------------------------------------------------------------------------- |
| 2 | |
| 3 | * GTSAM Copyright 2010, Georgia Tech Research Corporation, |
| 4 | * Atlanta, Georgia 30332-0415 |
| 5 | * All Rights Reserved |
| 6 | * Authors: Frank Dellaert, et al. (see THANKS for the full author list) |
| 7 | |
| 8 | * See LICENSE for the license information |
| 9 | |
| 10 | * -------------------------------------------------------------------------- */ |
| 11 | |
| 12 | /** |
| 13 | * @file METISOrdering.cpp |
| 14 | * @brief Simple robot motion example, with prior and two odometry measurements, |
| 15 | * using a METIS ordering |
| 16 | * @author Frank Dellaert |
| 17 | * @author Andrew Melim |
| 18 | */ |
| 19 | |
| 20 | /** |
| 21 | * Example of a simple 2D localization example optimized using METIS ordering |
| 22 | * - For more details on the full optimization pipeline, see OdometryExample.cpp |
| 23 | */ |
| 24 | |
| 25 | #include <gtsam/geometry/Pose2.h> |
| 26 | #include <gtsam/nonlinear/LevenbergMarquardtOptimizer.h> |
| 27 | #include <gtsam/nonlinear/Marginals.h> |
| 28 | #include <gtsam/nonlinear/NonlinearFactorGraph.h> |
| 29 | #include <gtsam/nonlinear/Values.h> |
| 30 | #include <gtsam/slam/BetweenFactor.h> |
| 31 | |
| 32 | using namespace std; |
| 33 | using namespace gtsam; |
| 34 | |
| 35 | int main(int argc, char** argv) { |
| 36 | NonlinearFactorGraph graph; |
| 37 | |
| 38 | Pose2 priorMean(0.0, 0.0, 0.0); // prior at origin |
| 39 | auto priorNoise = noiseModel::Diagonal::Sigmas(sigmas: Vector3(0.3, 0.3, 0.1)); |
| 40 | graph.addPrior(key: 1, prior: priorMean, model: priorNoise); |
| 41 | |
| 42 | Pose2 odometry(2.0, 0.0, 0.0); |
| 43 | auto odometryNoise = noiseModel::Diagonal::Sigmas(sigmas: Vector3(0.2, 0.2, 0.1)); |
| 44 | graph.emplace_shared<BetweenFactor<Pose2> >(args: 1, args: 2, args&: odometry, args&: odometryNoise); |
| 45 | graph.emplace_shared<BetweenFactor<Pose2> >(args: 2, args: 3, args&: odometry, args&: odometryNoise); |
| 46 | graph.print(str: "\nFactor Graph:\n" ); // print |
| 47 | |
| 48 | Values initial; |
| 49 | initial.insert(j: 1, val: Pose2(0.5, 0.0, 0.2)); |
| 50 | initial.insert(j: 2, val: Pose2(2.3, 0.1, -0.2)); |
| 51 | initial.insert(j: 3, val: Pose2(4.1, 0.1, 0.1)); |
| 52 | initial.print(str: "\nInitial Estimate:\n" ); // print |
| 53 | |
| 54 | // optimize using Levenberg-Marquardt optimization |
| 55 | LevenbergMarquardtParams params; |
| 56 | // In order to specify the ordering type, we need to se the "orderingType". By |
| 57 | // default this parameter is set to OrderingType::COLAMD |
| 58 | params.orderingType = Ordering::METIS; |
| 59 | LevenbergMarquardtOptimizer optimizer(graph, initial, params); |
| 60 | Values result = optimizer.optimize(); |
| 61 | result.print(str: "Final Result:\n" ); |
| 62 | |
| 63 | return 0; |
| 64 | } |
| 65 | |