This repository contains the code and necessary files to build and run the Pedestrian Audio Wearable System (PAWS) smartphone applications.
PAWS is a low-power connected system for improving pedestrian safety in current and future smart cities. PAWS uses microphones embedded into a headset combined with low-power feature extraction, signal processing, and machine learning, for detecting, localizing, and providing alerts of oncoming vehicles to pedestrians in noisy environments.
The contents of this repository are summarized below.
To reproduce the PAWS system, you must first install the PAWS smartphone application; details on how to do this can be found in the PAWS/Application/ folder.
Next, you must reproduce the PAWS front-end PCB. Details on how to do this can be found in the PAWS-FrontEnd repository.
A manual detailing how to use the smartphone application can be found in PAWS/PAWS_Manual.pdf. This file details basic app usage and how to train your own car detection models.
To reproduce the PAWS LE system, you must first install the PAWS LE smartphone application; details on how to do this can be found in the PAWS_LE/Application/ folder.
Next, you must reproduce the PAWS LE front-end PCB. Details on how to do this can be found in the PAWS-FrontEnd repository.
A manual detailing how to use the smartphone application can be found in PAWS_LE/PAWSLE_Manual.pdf. This file details basic app usage and how to train your own car detection models.
To learn more about PAWS, please visit our project page, or contact us at: stephen.xia@columbia.edu.
This repository is part of the Pedestrian Audio Wearable System (PAWS) project of the Intelligent and Connected Systems Lab (ICSL), Columbia University. For more information about our latest projects, please visit our group website.