Outcome Report

In medical applications and systems, the sensing, networking, and control functionalities of medical systems are often tightly coupled with the medical context, such as the patient’s physiological state and caregiver’s workflow. As much of that context information becomes available in real-time via connected medical devices and sensors, we are facing an urgent challenge to investigate and integrate context into network and security control under safety requirements, which is critical to the safety of patients in many applications. To accomplish this goal, this work aims at (i) establishing fundamental models for network and access control that incorporate context information, (ii) proposing networked solutions for medical applications under their safety and reliability requirements with a set of enabling new technologies, and (iii) designing, implementing, and deploying an open-source medical device and sensor network as a reference implementation to evaluate our designs in real-world scenarios. 

The main outcomes of this project are summarized into the following categories. 

Medical applications and systems

We built technologies on multiple medical and health domains. The first domain is Artificial Pancreas (AP), we have developed many techniques, including meal and exercise detection, data-driven blood glucose control based on patient’s historical records, and a machine learning based control based on MPC-guided policy search.  Our work contributes to the design of smart and fully closed loop insulin pumps, which have the potential to improve type 1 diabetes treatment and patient management. 

We built novel monitoring systems for manual toothbrushes and electrical toothbrushes. These systems achieve accurate detection of toothbrushing coverage of 16 different teeth surfaces, outperforming the state-of-the-art solutions in terms of the detection granularity and accuracy. If adopted by the industry, they could improve home oral hygiene effectiveness.  

We also built assistant tools for depression and bipolar disease diagnosis and detection using wearables. Some of our research results are used in clinical studies and gain good results. 

Wireless 

We worked on networking solutions that predict and optimize wireless networking reliability and efficiency. Our unique perspective is to utilize a user’s mobility and activity information as context to design highly efficient networking protocols and highly accurate networking controls.  

Developing and deploying the sensor for healthcare will help pervasive health monitoring in both hospital and home settings. Our work contributes to the algorithm design for mobile charger path planning for sensor networks. We study the problem of finding the charging path and maximizing the number of nodes charged within a fixed time horizon. We prove that this problem is APX-hard. By recursively decomposing the problem into sub-problems of searching sub-paths, we design a quasi-polynomial time algorithm that achieves polylogarithmic approximation to the optimum charging path. We also formally define the time-constrained data harvesting problem, which seeks an optimal data harvesting path in a network to collect as much data as possible within a time duration. We first characterize the performance bound given by the optimal data harvesting algorithm and show that the optimal algorithm significantly outperforms the random algorithm. Motivated by the theoretical analysis and proving the NP-completeness of the time-constrained data harvesting problem, we devise polynomial-time approximation schemes (PTAS) and mathematically prove the output being a constant-factor approximation of the optimal solution.

Our work on WiFi-based sensing focuses on multiple scenarios: indoor intrusion detection and outdoor user tracking. By fusing WiFi and visual signatures, we also built a system called RFCam to identify and track users in crowded public areas.

Control and Coordination

We investigated control problems of systems of systems for smart cities and smart transportation systems. Specifically, a decentralized conflict resolution framework for smart services is designed. As various smart services are increasingly deployed in modern cities, many unexpected conflicts arise due to various physical world couplings. Existing solutions for conflict resolution often rely on centralized control to enforce predetermined and fixed priorities of different services, which is challenging due to the inconsistent and private objectives of the services. Also, the centralized solutions miss opportunities to more effectively resolve conflicts according to the spatiotemporal locality of the conflicts. To address this issue, we designed a decentralized negotiation and conflict resolution framework named DeResolver, which allows services to resolve conflicts by communicating and negotiating with each other to reach a Pareto-optimal agreement autonomously and efficiently. Our design features a two-level semi-supervised learning-based algorithm to predict acceptable proposals and their rankings of each opponent through the negotiation.

A number of our works focus on smart transportation systems. We have studied the problems of model predictive taxi dispatch, coordination of multimodal public transportation systems, smart parking with real-time smart meter data, taxi games, and integration of an EV fleet and the power grid.

Last Modified: 12/04/2023
Modified by: Shan Lin