FHWS/IPIN2017
Archived
3
0
Fork 0
Code Issues 4 Pull Requests Releases Wiki Activity

eval final version jopooooooooo

Browse Source
This commit is contained in:
toni
2017-05-03 16:47:21 +02:00
parent ef74eafb60
commit b4720ae288
5 changed files with 129 additions and 76 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
code/filter/KLB.h
View File

@@ -212,6 +212,7 @@ struct ModeProbabilityTransitionNormal : public K::MarkovTransitionProbability<M
//create the matrix //create the matrix
Eigen::MatrixXd m(2,2); Eigen::MatrixXd m(2,2);
m << expKld, 1.0 - expKld, 1 - qualityWifi, qualityWifi; m << expKld, 1.0 - expKld, 1 - qualityWifi, qualityWifi;
//m << expKld, 1.0 - expKld, 0.00000000000000000001, 0.99999999999999999999;
return m; return m;

10
code/filter/Logic.h
View File

@@ -175,13 +175,13 @@ struct PFTransSimple : public K::ParticleFilterTransition<MyState, MyControl>{
p.state.position = newPosition; p.state.position = newPosition;
}else{ }else{
//no new position! //no new position!
#pragma omp atomic //#pragma omp atomic
noNewPositionCounter++; //noNewPositionCounter++;
} }
} }
std::cout << noNewPositionCounter << std::endl; //std::cout << noNewPositionCounter << std::endl;
} }
}; };
@@ -323,8 +323,8 @@ struct PFEval : public K::ParticleFilterEvaluation<MyState, MyObs> {
// Point3 posOld_m = p.state.positionOld.inMeter(); // Point3 posOld_m = p.state.positionOld.inMeter();
const double pWifi = getWIFI(observation, wifiObs, p.state.position); const double pWifi = getWIFI(observation, wifiObs, p.state.position);
const double pBaroPressure = getStairProb(p, observation.activity); //const double pBaroPressure = getStairProb(p, observation.activity);
//const double pBaroPressure = getBaroPressure(observation, p.state.relativePressure); const double pBaroPressure = getBaroPressure(observation, p.state.relativePressure);
//const double pBeacon = getBEACON(observation, p.state.position); //const double pBeacon = getBEACON(observation, p.state.position);
//small checks //small checks

2
code/filter/Structs.h
View File

@@ -22,7 +22,7 @@
struct MyState : public WalkState, public WalkStateHeading, public WalkStateSpread, public WalkStateFavorZ { struct MyState : public WalkState, public WalkStateHeading, public WalkStateSpread, public WalkStateFavorZ {
static Floorplan::IndoorMap* map; //static Floorplan::IndoorMap* map;
float relativePressure = 0.0f; float relativePressure = 0.0f;

147
code/main.cpp
View File

@@ -76,7 +76,7 @@ struct Data {
DataSetup IPIN2017 = { DataSetup IPIN2017 = {
mapDir + "SHL38.xml", mapDir + "SHL39.xml",
{ {
dataDir + "ipin2017/nogps/i-building/path1/1489769326868.csv", dataDir + "ipin2017/nogps/i-building/path1/1489769326868.csv",
@@ -96,22 +96,22 @@ struct Data {
mapDir + "wifi_fp_all.dat", mapDir + "wifi_fp_all.dat",
40, 40,
VAPGrouper::Mode::LAST_MAC_DIGIT_TO_ZERO, VAPGrouper::Mode::LAST_MAC_DIGIT_TO_ZERO,
mapDir + "grid_SHL38.dat" mapDir + "grid_SHL39.dat"
}; };
} data; } data;
Floorplan::IndoorMap* MyState::map; //Floorplan::IndoorMap* MyState::map;
void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPath) { K::Statistics<float> run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPath) {
std::vector<double> kld_data; std::vector<double> kld_data;
std::vector<double> quality_data; std::vector<double> quality_data;
// load the floorplan // load the floorplan
Floorplan::IndoorMap* map = Floorplan::Reader::readFromFile(setup.map); Floorplan::IndoorMap* map = Floorplan::Reader::readFromFile(setup.map);
MyState::map = map; //MyState::map = map;
WiFiModelLogDistCeiling WiFiModel(map); WiFiModelLogDistCeiling WiFiModel(map);
WiFiModel.loadAPs(map, Settings::WiFiModel::TXP, Settings::WiFiModel::EXP, Settings::WiFiModel::WAF); WiFiModel.loadAPs(map, Settings::WiFiModel::TXP, Settings::WiFiModel::EXP, Settings::WiFiModel::WAF);
@@ -121,7 +121,6 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
beaconModel.loadBeaconsFromMap(map, Settings::BeaconModel::TXP, Settings::BeaconModel::EXP, Settings::BeaconModel::WAF); beaconModel.loadBeaconsFromMap(map, Settings::BeaconModel::TXP, Settings::BeaconModel::EXP, Settings::BeaconModel::WAF);
//Assert::isFalse(beaconModel.getAllBeacons().empty(), "no Beacons stored within the map.xml"); //Assert::isFalse(beaconModel.getAllBeacons().empty(), "no Beacons stored within the map.xml");
// build the grid // build the grid
std::ifstream inp(setup.grid, std::ifstream::binary); std::ifstream inp(setup.grid, std::ifstream::binary);
Grid<MyNode> grid(20); Grid<MyNode> grid(20);
@@ -132,18 +131,15 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
onp.open(setup.grid); onp.open(setup.grid);
GridFactory<MyNode> factory(grid); GridFactory<MyNode> factory(grid);
factory.build(map); factory.build(map);
Importance::addImportance(grid);
WiFiGridEstimator::estimate(grid, WiFiModel, Settings::smartphoneAboveGround);
grid.write(onp); grid.write(onp);
} else { } else {
grid.read(inp); grid.read(inp);
} }
// add node-importance
Importance::addImportance(grid);
// stamp WiFi signal-strengths onto the grid
WiFiGridEstimator::estimate(grid, WiFiModel, Settings::smartphoneAboveGround);
// reading file // reading file
Offline::FileReader fr(setup.training[numFile]); Offline::FileReader fr(setup.training[numFile]);
@@ -170,9 +166,8 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
//std::shared_ptr<K::ParticleFilterInitializer<MyState>> init(new PFInitFixed(grid, GridPoint(1120.0f, 750.0f, 740.0f), 90.0f)); //std::shared_ptr<K::ParticleFilterInitializer<MyState>> init(new PFInitFixed(grid, GridPoint(1120.0f, 750.0f, 740.0f), 90.0f));
// mode 1 // mode 1
//std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode1(new PFInit(grid, 1)); std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode1(new PFInit(grid, 1));
std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode1(new PFInitFixed(grid, GridPoint(1120.0f, 750.0f, 740.0f), 90.0f, 1)); //std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode1(new PFInitFixed(grid, GridPoint(1120.0f, 750.0f, 740.0f), 90.0f, 1));
K::ParticleFilterMixing<MyState, MyControl, MyObs> mode1(Settings::numParticles, initMode1, Settings::Mode1::modeProbability); K::ParticleFilterMixing<MyState, MyControl, MyObs> mode1(Settings::numParticles, initMode1, Settings::Mode1::modeProbability);
mode1.setTransition(std::shared_ptr<PFTrans>(new PFTrans(grid, &ctrl))); mode1.setTransition(std::shared_ptr<PFTrans>(new PFTrans(grid, &ctrl)));
mode1.setEvaluation(std::shared_ptr<PFEval>(new PFEval(WiFiModel, beaconModel, grid))); mode1.setEvaluation(std::shared_ptr<PFEval>(new PFEval(WiFiModel, beaconModel, grid)));
@@ -183,8 +178,8 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
modes.push_back(mode1); modes.push_back(mode1);
// mode 2 // mode 2
//std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode2(new PFInit(grid, 2)); std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode2(new PFInit(grid, 2));
std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode2(new PFInitFixed(grid, GridPoint(1120.0f, 750.0f, 740.0f), 90.0f, 2)); //std::shared_ptr<K::ParticleFilterInitializer<MyState>> initMode2(new PFInitFixed(grid, GridPoint(1120.0f, 750.0f, 740.0f), 90.0f, 2));
K::ParticleFilterMixing<MyState, MyControl, MyObs> mode2(Settings::numParticles, initMode2, Settings::Mode2::modeProbability); K::ParticleFilterMixing<MyState, MyControl, MyObs> mode2(Settings::numParticles, initMode2, Settings::Mode2::modeProbability);
mode2.setTransition(std::shared_ptr<PFTransSimple>(new PFTransSimple(grid))); mode2.setTransition(std::shared_ptr<PFTransSimple>(new PFTransSimple(grid)));
mode2.setEvaluation(std::shared_ptr<PFEval>(new PFEval(WiFiModel, beaconModel, grid))); mode2.setEvaluation(std::shared_ptr<PFEval>(new PFEval(WiFiModel, beaconModel, grid)));
@@ -201,8 +196,8 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
K::InteractingMultipleModelParticleFilter<MyState, MyControl, MyObs> IMMAPF(modes, transitionProbabilityMatrix); K::InteractingMultipleModelParticleFilter<MyState, MyControl, MyObs> IMMAPF(modes, transitionProbabilityMatrix);
IMMAPF.setMixingSampler(std::unique_ptr<K::MixingSamplerDivergency<MyState, MyControl, MyObs>>(new K::MixingSamplerDivergency<MyState, MyControl, MyObs>())); IMMAPF.setMixingSampler(std::unique_ptr<K::MixingSamplerDivergency<MyState, MyControl, MyObs>>(new K::MixingSamplerDivergency<MyState, MyControl, MyObs>()));
IMMAPF.setJointEstimation(std::unique_ptr<K::JointEstimationPosteriorOnly<MyState, MyControl, MyObs>>(new K::JointEstimationPosteriorOnly<MyState, MyControl, MyObs>())); IMMAPF.setJointEstimation(std::unique_ptr<K::JointEstimationPosteriorOnly<MyState, MyControl, MyObs>>(new K::JointEstimationPosteriorOnly<MyState, MyControl, MyObs>()));
IMMAPF.setMarkovTransitionProbability(std::unique_ptr<ModeProbabilityTransition>(new ModeProbabilityTransition(grid, Settings::Mixing::lambda))); //IMMAPF.setMarkovTransitionProbability(std::unique_ptr<ModeProbabilityTransition>(new ModeProbabilityTransition(grid, Settings::Mixing::lambda)));
//IMMAPF.setMarkovTransitionProbability(std::unique_ptr<ModeProbabilityTransitionNormal>(new ModeProbabilityTransitionNormal(Settings::Mixing::lambda))); IMMAPF.setMarkovTransitionProbability(std::unique_ptr<ModeProbabilityTransitionNormal>(new ModeProbabilityTransitionNormal(Settings::Mixing::lambda)));
Timestamp lastTimestamp = Timestamp::fromMS(0); Timestamp lastTimestamp = Timestamp::fromMS(0);
@@ -216,10 +211,9 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
K::Statistics<float> errorStats; K::Statistics<float> errorStats;
//file writing for error data //file writing for error data
const long int t = static_cast<long int>(time(NULL)); const long int t = static_cast<long int>(time(NULL));
const std::string evalDir = errorDir + std::to_string(t); const std::string evalDir = errorDir + "final_" + std::to_string(t);
if(mkdir(evalDir.c_str(), S_IRWXU | S_IRWXG | S_IROTH | S_IXOTH) == -1){ if(mkdir(evalDir.c_str(), S_IRWXU | S_IRWXG | S_IROTH | S_IXOTH) == -1){
Assert::doThrow("Eval folder couldn't be created!"); Assert::doThrow("Eval folder couldn't be created!");
} }
@@ -227,6 +221,12 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
std::ofstream errorFile; std::ofstream errorFile;
errorFile.open (evalDir + "/" + std::to_string(numFile) + "_" + std::to_string(t) + ".csv"); errorFile.open (evalDir + "/" + std::to_string(numFile) + "_" + std::to_string(t) + ".csv");
std::ofstream kldFile;
kldFile.open (evalDir + "/kld_" + std::to_string(numFile) + "_" + std::to_string(t) + ".csv");
std::ofstream wifiFile;
wifiFile.open (evalDir + "/wifi_" + std::to_string(numFile) + "_" + std::to_string(t) + ".csv");
// parse each sensor-value within the offline data // parse each sensor-value within the offline data
for (const Offline::Entry& e : fr.getEntries()) { for (const Offline::Entry& e : fr.getEntries()) {
@@ -297,30 +297,30 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
//turn angle plot //turn angle plot
static float angleSumTurn = 0; angleSumTurn += ctrl.turnSinceLastTransition_rad; static float angleSumTurn = 0; angleSumTurn += ctrl.turnSinceLastTransition_rad;
plot.showAngle(1, angleSumTurn + M_PI, Point2(0.9, 0.9), "Turn: "); //plot.showAngle(1, angleSumTurn + M_PI, Point2(0.9, 0.9), "Turn: ");
//motion angle plot //motion angle plot
static float angleSumMotion = 0; angleSumMotion += ctrl.motionDeltaAngle_rad; static float angleSumMotion = 0; angleSumMotion += ctrl.motionDeltaAngle_rad;
plot.showAngle(2, angleSumMotion + M_PI, Point2(0.9, 0.8), "Motion: "); //plot.showAngle(2, angleSumMotion + M_PI, Point2(0.9, 0.8), "Motion: ");
plot.setEst(estPos); //plot.setEst(estPos);
plot.setGT(gtPos); //plot.setGT(gtPos);
plot.addParticles1(IMMAPF.getModes()[0].getParticles()); //plot.addParticles1(IMMAPF.getModes()[0].getParticles());
plot.addParticles2(IMMAPF.getModes()[1].getParticles()); //plot.addParticles2(IMMAPF.getModes()[1].getParticles());
plot.addEstimationNode(estPos); plot.addEstimationNode(estPos);
plot.addEstimationNodeSmoothed(IMMAPF.getModes()[1].getEstimation().position.inMeter()); //plot.addEstimationNodeSmoothed(IMMAPF.getModes()[1].getEstimation().position.inMeter());
//plot.gp << "set arrow 919 from " << tt.pos.x << "," << tt.pos.y << "," << tt.pos.z << " to "<< tt.pos.x << "," << tt.pos.y << "," << tt.pos.z+1 << "lw 3\n"; //plot.gp << "set arrow 919 from " << tt.pos.x << "," << tt.pos.y << "," << tt.pos.z << " to "<< tt.pos.x << "," << tt.pos.y << "," << tt.pos.z+1 << "lw 3\n";
//plot.gp << "set label 1001 at screen 0.02, 0.98 'base:" << relBaro.getBaseAvg() << " sigma:" << relBaro.getSigma() << " cur:" << relBaro.getPressureRealtiveToStart() << " hPa " << -relBaro.getPressureRealtiveToStart()/0.10/4.0f << " floor'\n"; //plot.gp << "set label 1001 at screen 0.02, 0.98 'base:" << relBaro.getBaseAvg() << " sigma:" << relBaro.getSigma() << " cur:" << relBaro.getPressureRealtiveToStart() << " hPa " << -relBaro.getPressureRealtiveToStart()/0.10/4.0f << " floor'\n";
int minutes = static_cast<int>(ts.sec()) / 60; //int minutes = static_cast<int>(ts.sec()) / 60;
plot.gp << "set label 1002 at screen 0.02, 0.94 'Time: " << minutes << ":" << static_cast<int>(static_cast<int>(ts.sec())%60) << "'\n"; //plot.gp << "set label 1002 at screen 0.02, 0.94 'Time: " << minutes << ":" << static_cast<int>(static_cast<int>(ts.sec())%60) << "'\n";
plot.gp << "set label 1003 at screen 0.02, 0.92 'KLD: " << ":" << kld_data.back() << "'\n"; //plot.gp << "set label 1003 at screen 0.02, 0.92 'KLD: " << ":" << kld_data.back() << "'\n";
plot.gp << "set label 1004 at screen 0.90, 0.98 'act:" << obs.activity << "'\n"; //plot.gp << "set label 1004 at screen 0.90, 0.98 'act:" << obs.activity << "'\n";
plot.gp << "set label 1011 at screen 0.90, 0.10 'Wifi Quality:" << __QUALITY << "'\n"; //plot.gp << "set label 1011 at screen 0.90, 0.10 'Wifi Quality:" << __QUALITY << "'\n";
plot.gp << "set label 1005 at screen 0.90, 0.08 'Prob. Mode1:" << IMMAPF.getModes()[0].getModePosteriorProbability() << "'\n"; //plot.gp << "set label 1005 at screen 0.90, 0.08 'Prob. Mode1:" << IMMAPF.getModes()[0].getModePosteriorProbability() << "'\n";
plot.gp << "set label 1006 at screen 0.90, 0.06 'Prob. Mode2:" << IMMAPF.getModes()[1].getModePosteriorProbability() << "'\n"; //plot.gp << "set label 1006 at screen 0.90, 0.06 'Prob. Mode2:" << IMMAPF.getModes()[1].getModePosteriorProbability() << "'\n";
int ones = 0; int ones = 0;
int twos = 0; int twos = 0;
@@ -342,16 +342,20 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
} }
} }
plot.gp << "set label 1007 at screen 0.90, 0.04 'Part. Mode1:" << ones << "'\n"; //plot.gp << "set label 1007 at screen 0.90, 0.04 'Part. Mode1:" << ones << "'\n";
plot.gp << "set label 1008 at screen 0.90, 0.02 'Part. Mode2:" << twos << "'\n"; //plot.gp << "set label 1008 at screen 0.90, 0.02 'Part. Mode2:" << twos << "'\n";
// error between GT and estimation // error between GT and estimation
float err_m = gtPos.getDistance(estPos); float err_m = gtPos.getDistance(estPos);
errorStats.add(err_m); errorStats.add(err_m);
errorFile << err_m << "\n"; errorFile << err_m << "\n";
plot.show(); kldFile << kld_data.back() << "\n";
usleep(10*10);
wifiFile << __QUALITY << "\n";
//plot.show();
//usleep(10*10);
lastTimestamp = ts; lastTimestamp = ts;
@@ -362,13 +366,15 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
} }
errorFile.close(); errorFile.close();
kldFile.close();
wifiFile.close();
std::cout << "Statistical Analysis: " << std::endl; std::cout << "Statistical Analysis: " << std::endl;
std::cout << "Median: " << errorStats.getMedian() << " Average: " << errorStats.getAvg() << std::endl; std::cout << "Median: " << errorStats.getMedian() << " Average: " << errorStats.getAvg() << " Std: " << errorStats.getStdDev() << std::endl;
//Write the current plotti buffer into file //Write the current plotti buffer into file
std::ofstream plotFile; std::ofstream plotFile;
plotFile.open(errorDir + std::to_string(numFile) + "_" + std::to_string(t) + ".gp"); plotFile.open(evalDir + "/gnuplot_" + std::to_string(numFile) + "_" + std::to_string(t) + ".gp");
plot.saveToFile(plotFile); plot.saveToFile(plotFile);
plotFile.close(); plotFile.close();
@@ -435,6 +441,8 @@ void run(DataSetup setup, int numFile, std::string folder, std::vector<int> gtPa
std::cout << "finished" << std::endl; std::cout << "finished" << std::endl;
sleep(1); sleep(1);
return errorStats;
} }
int main(int argc, char** argv) { int main(int argc, char** argv) {
@@ -442,23 +450,56 @@ int main(int argc, char** argv) {
//Testing files //Testing files
//run(data.BERKWERK, 6, "EVALBERGWERK"); // Nexus vor //run(data.BERKWERK, 6, "EVALBERGWERK"); // Nexus vor
//for(int i = 0; i < 5; ++i){ K::Statistics<float> statsAVG;
K::Statistics<float> statsMedian;
K::Statistics<float> statsSTD;
K::Statistics<float> statsQuantil;
K::Statistics<float> tmp;
for(int i = 0; i < 25; ++i){
// Settings::useKLB = false; // Settings::useKLB = false;
// //run(data.IPIN2017, 0, "ipin2017", Settings::Paths_IPIN2017::path1); // errorPair = run(data.IPIN2017, 1, "ipin2017", Settings::Paths_IPIN2017::path1);
// run(data.IPIN2017, 1, "ipin2017", Settings::Paths_IPIN2017::path1); // run(data.IPIN2017, 0, "ipin2017", Settings::Paths_IPIN2017::path1);
// run(data.IPIN2017, 2, "ipin2017", Settings::Paths_IPIN2017::path2); // errorPair = run(data.IPIN2017, 2, "ipin2017", Settings::Paths_IPIN2017::path2);
// //run(data.IPIN2017, 3, "ipin2017", Settings::Paths_IPIN2017::path2); // //run(data.IPIN2017, 3, "ipin2017", Settings::Paths_IPIN2017::path2);
// run(data.IPIN2017, 5, "ipin2017", Settings::Paths_IPIN2017::path3); // run(data.IPIN2017, 5, "ipin2017", Settings::Paths_IPIN2017::path3);
// //run(data.IPIN2017, 4, "ipin2017", Settings::Paths_IPIN2017::path3); // //run(data.IPIN2017, 4, "ipin2017", Settings::Paths_IPIN2017::path3);
//run(data.IPIN2017, 0, "ipin2017", Settings::Paths_IPIN2017::path1); tmp = run(data.IPIN2017, 4, "ipin2017", Settings::Paths_IPIN2017::path3);
statsMedian.add(tmp.getMedian());
statsAVG.add(tmp.getAvg());
statsSTD.add(tmp.getStdDev());
statsQuantil.add(tmp.getQuantile(0.75));
run(data.IPIN2017, 5, "ipin2017", Settings::Paths_IPIN2017::path3); tmp = run(data.IPIN2017, 5, "ipin2017", Settings::Paths_IPIN2017::path3);
//run(data.IPIN2017, 1, "ipin2017", Settings::Paths_IPIN2017::path1); statsMedian.add(tmp.getMedian());
//run(data.IPIN2017, 2, "ipin2017", Settings::Paths_IPIN2017::path2); statsAVG.add(tmp.getAvg());
//run(data.IPIN2017, 3, "ipin2017", Settings::Paths_IPIN2017::path2); statsSTD.add(tmp.getStdDev());
statsQuantil.add(tmp.getQuantile(0.75));
//run(data.IPIN2017, 4, "ipin2017", Settings::Paths_IPIN2017::path3); std::cout << "Iteration " << i << " completed" << std::endl;;
//} }
std::cout << "==========================================================" << std::endl;
std::cout << "Average of all statistical data: " << std::endl;
std::cout << "Median: " << statsMedian.getAvg() << std::endl;
std::cout << "Average: " << statsAVG.getAvg() << std::endl;
std::cout << "Standard Deviation: " << statsSTD.getAvg() << std::endl;
std::cout << "75 Quantil: " << statsQuantil.getAvg() << std::endl;
std::cout << "==========================================================" << std::endl;
//EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS
std::ofstream finalStatisticFile;
finalStatisticFile.open (errorDir + "/finalResults_Path3_AbsBaro.csv");
finalStatisticFile << "Average of all statistical data: \n";
finalStatisticFile << "Median: " << statsMedian.getAvg() << "\n";
finalStatisticFile << "Average: " << statsAVG.getAvg() << "\n";
finalStatisticFile << "Standard Deviation: " << statsSTD.getAvg() << "\n";
finalStatisticFile << "75 Quantil: " << statsQuantil.getAvg() << "\n";
finalStatisticFile.close();
//EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS EDIT THIS
return 0;
} }

45
tex/chapters/experiments.tex
View File

@@ -1,5 +1,7 @@
\section{Experiments} \section{Experiments}
% allgemeine infos über pfade und gebäude. wo
% bild: mit pfaden drauf und eventl. wifi qualität in jeweiligen bereichen? (kann frank das)
All upcoming experiments were carried out on four floors of a \SI{77}{m} x \SI{55}{m} sized faculty building. All upcoming experiments were carried out on four floors of a \SI{77}{m} x \SI{55}{m} sized faculty building.
It includes several staircases and elevators and has a ceiling height of about \SI{3}{m}. It includes several staircases and elevators and has a ceiling height of about \SI{3}{m}.
Nevertheless, the grid was generated for the complete campus and thus outdoor areas like the courtyard are also walkable. Nevertheless, the grid was generated for the complete campus and thus outdoor areas like the courtyard are also walkable.
@@ -11,40 +13,49 @@ As mentioned before, we omit any time-consuming calibration processes and use th
The position of the access-points (about five per floor) is known beforehand. The position of the access-points (about five per floor) is known beforehand.
Due to legal terms, we are not allowed to depict their positions and therefore omit this information within the figures. Due to legal terms, we are not allowed to depict their positions and therefore omit this information within the figures.
% gewählte parameter (auch mal die optimieren wifi parameter testen)
We arranged three distinct walks (see also fig. \ref{}). We arranged three distinct walks (see also fig. \ref{}).
The measurements for the walks were recorded using a Motorola Nexus 6 at 2.4 GHz band only. The measurements for the walks were recorded using a Motorola Nexus 6 at 2.4 GHz band only.
The computation was done offline as described in algorithm \ref{}. The computation was done offline as described in algorithm \ref{}.
For each walk we deployed $xx$ MC runs using 5000 Particles. For each walk we deployed $xx$ MC runs using 5000 Particles for each mode.
Instead of an initial position and heading, all walks start with a uniform distribution (random position and heading) as prior. Instead of an initial position and heading, all walks start with a uniform distribution (random position and heading) as prior.
For the filtering we used $\sigma_\text{wifi} = 8.0$ as uncertainties, both growing with each measurement's age. For the filtering we used $\sigma_\text{wifi} = 8.0$ as uncertainties, both growing with each measurement's age.
While the pressure change was assumed to be \SI{0.105}{$\frac{\text{\hpa}}{\text{\meter}}$}, all other barometer-parameters are determined automatically. While the pressure change was assumed to be \SI{0.105}{$\frac{\text{\hpa}}{\text{\meter}}$}, all other barometer-parameters are determined automatically.
The step size $\mStepSize$ for the transition was configured to be \SI{70}{\centimeter} with an allowed derivation of \SI{10}{\percent}. The heading deviation was set to \SI{25}{\degree}. The step size $\mStepSize$ for the transition was configured to be \SI{70}{\centimeter} with an allowed derivation of \SI{10}{\percent}. The heading deviation was set to \SI{25}{\degree}.
KLD with normal dist and kernel density drawing from grid. % wie für die kld gezogen? begründen warum wir nun keine parzenschätzung machen (weil ähnliche ergebnisse)
To calculate \eqref{equ:KLD} and thus the Kullback-Leibler divergence, we need to sample densities from both modes likewise.
The grid is suitable for this purpose.
However, sampling at any vertex $\mVertexA$ of the grid, given just a set of random variables (particles), is not the easiest task.
We need to estimate the posterior distribution given by the respective particle sets.
A common way is to deploy a kernel density estimation using a Gaussian distribution as kernel.
The density of a specific point $\hat\mStateVec_{t} = \fPos{\mVertexA}$ is then given by
%
\begin{equation}
p(\hat\mStateVec_{t} \mid m_t, \mObsVec_{1:t}) = \sum_{i=1}^{N_{m_t}} \mathcal{N}(d^i_{\text{KL}} \mid 0, \sigma_{\text{KL}})
\enspace ,
\end{equation}
%
while $d^i_{\text{KL}}$ is the euclidean distance between the considered point's $\hat\mStateVec_{t}$ and all particles $\fPos{\vec{X}_t^{i,m_t}}$ of the mode. The variance $\sigma_{\text{KL}}$ is set to \SI{1}{m}.
It is well known, that the computation of the kernel density estimation is rather slow, thus we also used a much simpler estimation by assuming a multivariate Gaussian distribution for both modes.
Here, the mean is given by weighted arithmetic mean of the particles and the variance is defined by the sample covariance matrix.
% ground truth
The ground truth is measured by recording a timestamp at marked spots on the walking route. When passing a marker, the pedestrian clicked a button on the smartphone application. The ground truth is measured by recording a timestamp at marked spots on the walking route. When passing a marker, the pedestrian clicked a button on the smartphone application.
Between two consecutive points, a constant movement speed is assumed. Between two consecutive points, a constant movement speed is assumed.
Thus, the ground truth might not be \SI{100}{\percent} accurate, but fair enough for error measurements. Thus, the ground truth might not be \SI{100}{\percent} accurate, but fair enough for error measurements.
The approximation error is then calculated by comparing the interpolated ground truth position with the current estimation \cite{Fetzer2016OMC}. The approximation error is then calculated by comparing the interpolated ground truth position with the current estimation \cite{Fetzer2016OMC}.
% allgemeine infos über pfade und gebäude. wo
% bild: mit pfaden drauf und eventl. wifi qualität in jeweiligen bereichen? (kann frank das)
% gewählte parameter (auch mal die optimieren wifi parameter testen)
% wie für die kld gezogen? begründen warum wir nun keine parzenschätzung machen (weil ähnliche ergebnisse)
% ground truth
% maß für die streuung der verteilung (diversity von partikeln) % maß für die streuung der verteilung (diversity von partikeln)
error at the beginning always very high. about 44 meters. therefore the median is better value oder 75 quantil.
% zeigen das es stucken verhindert (eventl. hier eine andere aufnahme die mitten drinnen stecken bleibt) % zeigen das es stucken verhindert (eventl. hier eine andere aufnahme die mitten drinnen stecken bleibt)
% bild: stucken im raum + nicht mehr stucken im raum % bild: stucken im raum + nicht mehr stucken im raum + kld mit anzeigen
% zeigen das schlechtes wi-fi (zu hohe diversity) behoben wird. % zeigen das schlechtes wi-fi (zu hohe diversity) behoben wird.
% bild: lauf auf der rechten seite des gebäudes zeige mit und ohne wifi faktor (schlechtes wifi einzeichnen) % bild: lauf auf der rechten seite des gebäudes zeige mit und ohne wifi faktor (schlechtes wifi einzeichnen)