README.md

Learnable Spatio-Temporal Map Embeddings for Deep Inertial Localization

Dennis Melamed, Karnik Ram, Vivek Roy, Kris Kitani.

[arXiv] [Project Page]

In IROS 2022

Our proposed method utilizes learnable spatio-temporal map priors to reduce drift in inertial odometry.

Setup

conda env create --file map-prior.yml

Train

python main.py --mode TrainLightning --dataset BLE_IMU --building building2_f1 --train-gpus 2 --data-sample-rate 60
python main.py --mode TrainLightning --dataset IDOL --building building2_f1 --train-gpus 2 --data-sample-rate 100

Filter

python main.py --mode RunFilter --dataset BLE_IMU --building building2_f1
python main.py --mode RunFilter --dataset BLE_IMU --building building2_f1 --no-filter-allow-ble-update
python main.py --mode RunFilter --dataset IDOL --building building1 --filter-update-rate 100
python main.py --mode RunFilter --dataset IDOL --building building1 --filter-update-rate 100 --no-filter-allow-reinit