| 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 SmartRangeExample_plaza2.cpp |
| 14 | * @brief A 2D Range SLAM example |
| 15 | * @date June 20, 2013 |
| 16 | * @author FRank Dellaert |
| 17 | */ |
| 18 | |
| 19 | // Both relative poses and recovered trajectory poses will be stored as Pose2 objects |
| 20 | #include <gtsam/geometry/Pose2.h> |
| 21 | |
| 22 | // Each variable in the system (poses and landmarks) must be identified with a unique key. |
| 23 | // We can either use simple integer keys (1, 2, 3, ...) or symbols (X1, X2, L1). |
| 24 | // Here we will use Symbols |
| 25 | #include <gtsam/inference/Symbol.h> |
| 26 | |
| 27 | // We want to use iSAM2 to solve the range-SLAM problem incrementally |
| 28 | #include <gtsam/nonlinear/ISAM2.h> |
| 29 | |
| 30 | // iSAM2 requires as input a set set of new factors to be added stored in a factor graph, |
| 31 | // and initial guesses for any new variables used in the added factors |
| 32 | #include <gtsam/nonlinear/NonlinearFactorGraph.h> |
| 33 | #include <gtsam/nonlinear/Values.h> |
| 34 | |
| 35 | // We will use a non-liear solver to batch-inituialize from the first 150 frames |
| 36 | #include <gtsam/nonlinear/LevenbergMarquardtOptimizer.h> |
| 37 | |
| 38 | // In GTSAM, measurement functions are represented as 'factors'. Several common factors |
| 39 | // have been provided with the library for solving robotics SLAM problems. |
| 40 | #include <gtsam/slam/BetweenFactor.h> |
| 41 | #include <gtsam/sam/RangeFactor.h> |
| 42 | |
| 43 | // To find data files, we can use `findExampleDataFile`, declared here: |
| 44 | #include <gtsam/slam/dataset.h> |
| 45 | |
| 46 | // Standard headers, added last, so we know headers above work on their own |
| 47 | #include <fstream> |
| 48 | #include <iostream> |
| 49 | |
| 50 | using namespace std; |
| 51 | using namespace gtsam; |
| 52 | namespace NM = gtsam::noiseModel; |
| 53 | |
| 54 | // data available at http://www.frc.ri.cmu.edu/projects/emergencyresponse/RangeData/ |
| 55 | // Datafile format (from http://www.frc.ri.cmu.edu/projects/emergencyresponse/RangeData/log.html) |
| 56 | |
| 57 | // load the odometry |
| 58 | // DR: Odometry Input (delta distance traveled and delta heading change) |
| 59 | // Time (sec) Delta Dist. Trav. (m) Delta Heading (rad) |
| 60 | typedef pair<double, Pose2> TimedOdometry; |
| 61 | list<TimedOdometry> readOdometry() { |
| 62 | list<TimedOdometry> odometryList; |
| 63 | string drFile = findExampleDataFile(name: "Plaza2_DR.txt" ); |
| 64 | ifstream is(drFile); |
| 65 | if (!is) throw runtime_error("Plaza2_DR.txt file not found" ); |
| 66 | |
| 67 | while (is) { |
| 68 | double t, distance_traveled, delta_heading; |
| 69 | is >> t >> distance_traveled >> delta_heading; |
| 70 | odometryList.push_back( |
| 71 | x: TimedOdometry(t, Pose2(distance_traveled, 0, delta_heading))); |
| 72 | } |
| 73 | is.clear(); /* clears the end-of-file and error flags */ |
| 74 | return odometryList; |
| 75 | } |
| 76 | |
| 77 | // load the ranges from TD |
| 78 | // Time (sec) Sender / Antenna ID Receiver Node ID Range (m) |
| 79 | typedef std::tuple<double, size_t, double> RangeTriple; |
| 80 | vector<RangeTriple> readTriples() { |
| 81 | vector<RangeTriple> triples; |
| 82 | string tdFile = findExampleDataFile(name: "Plaza2_TD.txt" ); |
| 83 | ifstream is(tdFile); |
| 84 | if (!is) throw runtime_error("Plaza2_TD.txt file not found" ); |
| 85 | |
| 86 | while (is) { |
| 87 | double t, sender, receiver, range; |
| 88 | is >> t >> sender >> receiver >> range; |
| 89 | triples.push_back(x: RangeTriple(t, receiver, range)); |
| 90 | } |
| 91 | is.clear(); /* clears the end-of-file and error flags */ |
| 92 | return triples; |
| 93 | } |
| 94 | |
| 95 | // main |
| 96 | int main(int argc, char** argv) { |
| 97 | |
| 98 | // load Plaza1 data |
| 99 | list<TimedOdometry> odometry = readOdometry(); |
| 100 | // size_t M = odometry.size(); |
| 101 | |
| 102 | vector<RangeTriple> triples = readTriples(); |
| 103 | size_t K = triples.size(); |
| 104 | |
| 105 | // parameters |
| 106 | size_t incK = 50; // minimum number of range measurements to process after |
| 107 | bool robust = false; |
| 108 | |
| 109 | // Set Noise parameters |
| 110 | Vector priorSigmas = Vector3(0.01, 0.01, 0.01); |
| 111 | Vector odoSigmas = Vector3(0.05, 0.01, 0.2); |
| 112 | double sigmaR = 100; // range standard deviation |
| 113 | const NM::Base::shared_ptr // all same type |
| 114 | priorNoise = NM::Diagonal::Sigmas(sigmas: priorSigmas), //prior |
| 115 | odoNoise = NM::Diagonal::Sigmas(sigmas: odoSigmas), // odometry |
| 116 | gaussian = NM::Isotropic::Sigma(dim: 1, sigma: sigmaR), // non-robust |
| 117 | tukey = NM::Robust::Create(robust: NM::mEstimator::Tukey::Create(k: 15), noise: gaussian), //robust |
| 118 | rangeNoise = robust ? tukey : gaussian; |
| 119 | |
| 120 | // Initialize iSAM |
| 121 | ISAM2 isam; |
| 122 | |
| 123 | // Add prior on first pose |
| 124 | Pose2 pose0 = Pose2(-34.2086489999201, 45.3007639991120, -2.02108900000000); |
| 125 | NonlinearFactorGraph newFactors; |
| 126 | newFactors.addPrior(key: 0, prior: pose0, model: priorNoise); |
| 127 | |
| 128 | // initialize points (Gaussian) |
| 129 | Values initial; |
| 130 | initial.insert(j: symbol(c: 'L', j: 1), val: Point2(-68.9265, 18.3778)); |
| 131 | initial.insert(j: symbol(c: 'L', j: 6), val: Point2(-37.5805, 69.2278)); |
| 132 | initial.insert(j: symbol(c: 'L', j: 0), val: Point2(-33.6205, 26.9678)); |
| 133 | initial.insert(j: symbol(c: 'L', j: 5), val: Point2(1.7095, -5.8122)); |
| 134 | Values landmarkEstimates = initial; // copy landmarks |
| 135 | initial.insert(j: 0, val: pose0); |
| 136 | |
| 137 | // set some loop variables |
| 138 | size_t i = 1; // step counter |
| 139 | size_t k = 0; // range measurement counter |
| 140 | Pose2 lastPose = pose0; |
| 141 | size_t countK = 0; |
| 142 | |
| 143 | // Loop over odometry |
| 144 | gttic_(iSAM); |
| 145 | for(const TimedOdometry& timedOdometry: odometry) { |
| 146 | //--------------------------------- odometry loop ----------------------------------------- |
| 147 | double t; |
| 148 | Pose2 odometry; |
| 149 | std::tie(args&: t, args&: odometry) = timedOdometry; |
| 150 | |
| 151 | // add odometry factor |
| 152 | newFactors.push_back( |
| 153 | factor: BetweenFactor<Pose2>(i - 1, i, odometry, |
| 154 | NM::Diagonal::Sigmas(sigmas: odoSigmas))); |
| 155 | |
| 156 | // predict pose and add as initial estimate |
| 157 | Pose2 predictedPose = lastPose.compose(g: odometry); |
| 158 | lastPose = predictedPose; |
| 159 | initial.insert(j: i, val: predictedPose); |
| 160 | landmarkEstimates.insert(j: i, val: predictedPose); |
| 161 | |
| 162 | // Check if there are range factors to be added |
| 163 | while (k < K && t >= std::get<0>(t&: triples[k])) { |
| 164 | size_t j = std::get<1>(t&: triples[k]); |
| 165 | double range = std::get<2>(t&: triples[k]); |
| 166 | RangeFactor<Pose2, Point2> factor(i, symbol(c: 'L', j), range, rangeNoise); |
| 167 | // Throw out obvious outliers based on current landmark estimates |
| 168 | Vector error = factor.unwhitenedError(x: landmarkEstimates); |
| 169 | if (k <= 200 || std::abs(x: error[0]) < 5) |
| 170 | newFactors.push_back(factor); |
| 171 | k = k + 1; |
| 172 | countK = countK + 1; |
| 173 | } |
| 174 | |
| 175 | // Check whether to update iSAM 2 |
| 176 | if (countK > incK) { |
| 177 | gttic_(update); |
| 178 | isam.update(newFactors, newTheta: initial); |
| 179 | gttoc_(update); |
| 180 | gttic_(calculateEstimate); |
| 181 | Values result = isam.calculateEstimate(); |
| 182 | gttoc_(calculateEstimate); |
| 183 | lastPose = result.at<Pose2>(j: i); |
| 184 | // update landmark estimates |
| 185 | landmarkEstimates = Values(); |
| 186 | landmarkEstimates.insert(j: symbol(c: 'L', j: 1), val: result.at(j: symbol(c: 'L', j: 1))); |
| 187 | landmarkEstimates.insert(j: symbol(c: 'L', j: 6), val: result.at(j: symbol(c: 'L', j: 6))); |
| 188 | landmarkEstimates.insert(j: symbol(c: 'L', j: 0), val: result.at(j: symbol(c: 'L', j: 0))); |
| 189 | landmarkEstimates.insert(j: symbol(c: 'L', j: 5), val: result.at(j: symbol(c: 'L', j: 5))); |
| 190 | newFactors = NonlinearFactorGraph(); |
| 191 | initial = Values(); |
| 192 | countK = 0; |
| 193 | } |
| 194 | i += 1; |
| 195 | //--------------------------------- odometry loop ----------------------------------------- |
| 196 | } // end for |
| 197 | gttoc_(iSAM); |
| 198 | |
| 199 | // Print timings |
| 200 | tictoc_print_(); |
| 201 | |
| 202 | // Write result to file |
| 203 | Values result = isam.calculateEstimate(); |
| 204 | ofstream os2("rangeResultLM.txt" ); |
| 205 | for (const auto& [key, point] : result.extract<Point2>()) |
| 206 | os2 << key << "\t" << point.x() << "\t" << point.y() << "\t1" << endl; |
| 207 | ofstream os("rangeResult.txt" ); |
| 208 | for (const auto& [key, pose] : result.extract<Pose2>()) |
| 209 | os << key << "\t" << pose.x() << "\t" << pose.y() << "\t" << pose.theta() << endl; |
| 210 | exit(status: 0); |
| 211 | } |
| 212 | |
| 213 | |