This work is a part of On Attitude Estimation with Smartphones project
http://tyrex.inria.fr/mobile/benchmarks-attitude/
Abstract
We investigate the precision of attitude estimation algorithms in the particular context of pedestrian navigation with commodity smartphones and their inertial/magnetic sensors. We report on an extensive comparison and experimental analysis of existing algorithms. We focus on typical motions of smartphones when carried by pedestrians. We use a precise ground truth obtained from a motion capture system. We test state-of-the-art attitude estimation techniques with several smartphones, in the presence of magnetic perturbations typically found in buildings. We discuss the obtained results, analyze advantages and limits of current technologies for attitude estimation in this context. Furthermore, we propose a new technique for limiting the impact of magnetic perturbations with any attitude estimation algorithm used in this context. We show how our technique compares and improves over previous works.
How it works ?
To start, open a matlab console. Go to src folder. Add this folder and subfolders to the matlab path:
cd YOUR_PATH/benchmarks-attitude-smartphones/src/
loadProject;
Create a new benchmark instance. Load sensors noises, calibrations and datasets. Then start the benchmark process:
benchmark = AttitudeBenchmarks;
benchmark.load;
benchmark.process;
The benchmark process takes a while and memory if you compute all datasets with all algorithms, all sampling, all calibrations... (e.g. 26 hours on a macbook pro and 25 Go). In order to avoid to restart the full process each time, you can save/open the benchmark:
save(AttitudeBenchmarks.path, 'benchmark', '-v7.3'); % current state is saved in ./datasets/saved/ (default path)
load(AttitudeBenchmarks.path);
Results are stored in AttitudeBenchmarkResults class and aggregated data on means can be computed using compare function
results = benchmark.results;
results.compare('calibration');
results.compare('motions');
results.compare('motionsmag');
results.compare('mag');
results.compare('devices');
results.compare('sampling');
results.compare('processingTime');
Some stats on datasets can be computed:
benchmark.stats;
License
This project is under the CeCILL license.
Authors
Thibaud Michel
[email protected]
Hassen Fourati
Pierre Genevès
Nabil Layaïda
Tyrex Team, LIG, Inria (France), 2016
NEcS Team, Gipsa-Lab, Inria (France), 2016