Continuous sensing of a large pool of indoor activities at different scale (micro-scale and macro-scale body movements) pervasively through contactless RF sensing
Key Features
Uniqueness: MARS offers a unique radar-based system to contniously monitor a diverse set of human activities in a multi-user environment. The core idea is to combine mmWave sensing with an intelligent rotor-based 360-degree scanning for room-scale activity monitoring by periodically switching the sensing resolution. The core features of the platform are as follows.
- Diverse Activity Set: We have included a wide array of 19 different activity classes from Activities of Daily Living (ADLs), Instrumental Activities of Daily Living (IADLs), and daily indoor exercises – (i) macro activities like walking, running, jumping, clapping, lunges, squats, waving, vacuum cleaning, folding clothes, changing clothes, and (ii) micro activities like laptop-typing, phone-talking, phone-typing, sitting, playing guitar, eating food, combing hair, brushing teeth, and drinking water to ensure comprehensive coverage of daily human behaviors.
- Multiple Subjects: MARS uniquely includes multiple subjects performing various activities simultaneously, unlike traditional datasets that typically focus on single-subject activities. This feature allows our platform to account for the complexity and variability of real-world scenarios where interactions between multiple individuals are common.
- Adaptive Doppler Resolution: As we observe that the macro and micro-scale activities can be better captured by dynamically switching the radar sensitivity, MARS, therefore, employs an adaptive doppler resolution strategy to sense activities at different scales.
- Real-world Setups: The platform is tested under diverse scenarios of real indoor environments with the presence of furniture and other objects usually found in an indoor setup. This promotes the real-world applicability of the developed platform.
Licensing:The platform source and the collected dataset is free to download and can be used with GNU Affero General Public License for non-commercial purposes. All participants signed forms consenting to the collected mmwave dataset and activity labels for non-commercial research purposes. The institute’s ethical review committee has approved the field study (Order No: IIT/SRIC/DEAN/2023, Dated July 31, 2023).
Argha Sen
IIT Kharagpur, India
Swadhin Pradhan
Cisco Systems, US
Anirban Das
IIT Kharagpur, India
Sandip Chakraborty
IIT Kharagpur, India
Publications
- Argha Sen, Anirban Das, Swadhin Pradhan, and Sandip Chakraborty: "Continuous Multi-user Activity Tracking via Room-Scale mmWave Sensing", ACM/IEEE IPSN 2024
- Argha Sen, Anirban Das, Swadhin Pradhan, and Sandip Chakraborty: "Demo Abstract: MARS: An mmWave-based Multi-user Activity Tracking Solution", ACM/IEEE IPSN 2024 Demos
Funding and Support
For questions and general feedback, contact
Argha Sen