Research Projects @ USC
| Reconfigurable Body Area Sensor Network for Motor Rehabilitation - PAST PROJECT |
|---|
Project DescriptionAn increasingly diverse set of sensing technologies is appearing in smart products and research projects. The goal of this project is to develop an interactive and intelligent wearable computing platform using various sensors for motion analysis, health monitoring, sports medicine, and motor rehabilitation/physical therapy. We have been building up a real-time multimodal patient monitoring Hardware/Software system based on SunSPOT technology. Many kinds of sensors, such as Bend Sensor, Flex/Force sensor, 3-Axis Accelerometers/Gyroscopes (IMU), IR/Ultrasonic rangers are being integrated onto the SunSPOT platforms. A fault-tolerant, dynamically configured Wireless Body Sensor Network has been set up to collect data from nodes attached to different parts of the human body. In-situ real-time feedback is provided based on sensor inputs to guide patients for proper movements. Meanwhile, the software installed on the PC not only supports startup calibration, displaying streaming data from multiple nodes, but also supports command-based remote reconfiguration on on-body nodes over the air. The data of each patient is stored and archived in the backend database for future analysis. This prototype system has been integrated with 3D game engines to realize "Virtual Reality" and used by the physical therapists at Precision Rehabilitation Clinic to facilitate their work. You can find the demo video here on my Youtube channel.
Our prototype system also needs to be further improved in several aspects.
- A Weka machine learning engine is being integrated to support online activity recognition and inference.
- Compared to mote-class systems (Tmote, MICA2), SunSPOT has a more powerful processing unit (32-bit ARM9 microprocessor) and a much larger data storage (4Mbit Flash). This makes in-network processing and local data storage possible, and therefore reduces radio usage so as to reduce power consumption. In addition, we are going to embed lightweight classification algorithms into SunSPOTs to provide proper guidance to patients in realtime.
- We are quite interested in answering the question: "Which statistical machine learning algorithm is the best one for one specific application?". In order to solve this problem, we need to switch between different algorithms automatically, and the inference engine should be reconfigurable under the help of user feedback.
- Most body sensor network (BSN) systems are for individualized care. We are going to expand the system to enable multi-agent applications. Personal BSN system then can interact with the environment or other people within the inter-body network.
Affiliation
Dept of Electrical Engineering, USC. VRPsych Laboratory, USC. Institute for Creative Technologies, USC. Division of Biokinesiology and Physical Therapy, USC.
Presentation
- Introduction to Sensor-based System Design
Institute for Creative Technologies, USC - Sensor-based Interactive Technologies for Motor Rehabilitation
Institute for Creative Technologies, USC
- M. Zhang, A. Sawchuk. "A Customizable Framework of Body Area Sensor Network for Rehabilitation". In 2nd International Symposium on Applied Sciences in Biomedical and Communication Technologies (Isabel), November 2009.  [PDF]
| Compact representation of Markov Decision Process (MDP) framework for Body Area Sensor Network - PAST PROJECT |
|---|
Project DescriptionMarkov Decision Process (MDP) can be used as an optimal decision policies in power control. The learning of a small decision policy is key to deploying the model in small sensor nodes with limited memory. The decision process enables distributed sensor nodes to adapt their sampling rates in response to changing event criticality and the availability of resources (energy) at each node. The globally optimal policy is first calculated offline using an MDP and deployed onto each node. However, the space complexity of the representation is exponential in the number of the sensor nodes and discretization grain of the problem. We compare the capability of the compact representation of the optimal decision policy by using different base supervised learners. The results show that unpruned decision trees and high confidence pruned decision trees provide the lowest error rate while the required node number of the decision tree is enough small to be stored in the sensors. Ensembles of lower-confidence trees are capable of perfect representation with only an order of magnitude increase in classifier size compared to individual pruned trees.
Affiliation
Dept of Electrical Engineering, USC.
Publications
- S.P. Liu, M. Zhang. "Evaluation of Learning Algorithms for Optimal Policy Representation in Sensor-Network based Human Health Monitoring Systems". 7th International Conference on Information, Communications and Signal Processing (ICICS), December 2009. 
| Event-Triggered Indoor Localization using IEEE 802.15.4 Ad Hoc Network - PAST PROJECT |
|---|
Project DescriptionIn United States, Fall is responsible of 70% of accidental death in persons aged 75+. Therefore, an automatic fall detection system becomes a clear necessity. In the mean while, especially in the case of an emergency, an immediate on-site assistance should be provided to the right spot with a granularity beyond GPS locations. In this project, a wireless sensor network is established for indoor localization, which is triggered whenever a fall is detected. A SOS message is routed to the basestation (either a mobile gateway or an internet gateway) and is sent out via mobile infrastructure to inform both the relatives and the nearest medical center.
Affiliation
Dept of Electrical Engineering, USC. VRPsych Laboratory, USC. Institute for Creative Technologies, USC.
| OCRdroid: A Framework to Digitize Text on Smart Phones - PAST PROJECT |
|---|
Project DescriptionA s demand grows for mobile phone applications, research in optical character recognition, a technology well developed for scanned documents, is shifting focus to the recognition of text embedded in digital photographs. In this project, we present OCRdroid, a generic framework for developing OCR-based applications on mobile phones. We believe this mobile solution to extract information from physical world is a good match for future trend. OCRdroid combines a light-weight image preprocessing suite installed inside the mobile phone and an OCR engine connected to a backend server. We demonstrate the power and functionality of this framework by implementing two applications called PocketPal and PocketReader based on OCRdroid on HTC Android G1 mobile phone. Initial evaluations of these pilot experiments demonstrate the potential of using OCRdroid framework for realworld OCR-based mobile applications. For more information, please check the paper listed below and the project site with a demo video here.
Affiliation
Dept of Computer Science, USC. The Robotic Embedded Systems Laboratory, USC.
Publications
- M. Zhang, A. Joshi, R. Kadmawala, K. Dantu, S. Poduri, G. S. Sukhatme. "OCRdroid: A Framework to Digitize Text on Smart Phones". In The First Annual International Conference on Mobile Computing, Applications, and Services (MobiCASE), October 2009.  [PDF]
| Mobile Labor Market: A General Platform for Mobile Community - PAST PROJECT |
|---|
Project DescriptionMobile Labor Market (MLM) is a project under “ Mobile Voices ”, which is a collaboration between the Annenberg School for Communication at University of Southern California and the Institute of Popular Education of Southern California (IDEPSCA), a nonprofit institution serving low-income Latino immigrants in Southern California. This project aims to build a mobile platform for the research in mobile community. The platform is used by low-wage immigrants community in Los Angeles area, helping people without computers access to have greater participation in job huntings using their mobile phones. The GPRS-based multi-user software system adopts a Client-Server based 3-Tier architecture. All the information is password protected and ciphered for privacy. Data are updated in real time, and stored in the backend web-serviced database. A Nokia N95 smart phone is used acting as a basestation connected to the backend database. Local contractors and workers have corresponding programs installed on their mobile phones. A worker can upload and update his profile, search and apply for jobs which match his skills. A contractor can also upload and update his profile, post new jobs, search for job applicants, and pick up applicants for specific jobs. In addition, an evaluation system is introduced so that workers can be evaluated based on their performance. The prototype system triggers researchers’ interest at Nokia Research Center at Santa Monica, and now becomes the tool for Annenberg professors conducting their research.
Affiliation
Dept of Electrical Engineering, USC. Annenberg School for Communication, USC. IDEPSCA, CA.
Presentation
- Introduction to Mobile Labor Market - Utilizing the power of mobile devices  
Annenberg School for Communication, USC / Institute of Popular Education of Southern California
- M. Zhang, "Mobile Labor Market: A General Platform for Mobile Community"   [PDF]
| Parallel & High Performance Computing on Graphic Processor - PAST PROJECT |
|---|
Project DescriptionThe Graphics Processing Unit (GPU) has become an integral part of today’s mainstream computing systems. Over the past few years, there has been a marked increase in the performance and capabilities of GPUs. The modern GPU is not only a powerful graphics engine but also a highly parallel programmable processor featuring peak arithmetic and memory bandwidth that substantially outpaces its CPU counterpart. This project aims to investigate the compilation techniques for exploiting parallelism and the computational power of GPU based on Nvidia's CUDA technology. The stream architecture of GPU, and programming model of CUDA, including memory hierarchy, thread hierarchy and scheduling, are also what I am interested in. I have developed optimized parallel codes for large sparse matrix multiplication, Bitonic-Sort and Quicksort array sorting algorithms, and at last, compared the performance between Nvidia GeForce 8800 GTX GPU and Intel Multi-core CPU.
Affiliation
Information Sciences Institute, USC.
Presentation
- Parallel & high performance computing using Nvidia’s CUDA  
Information Sciences Institute, Computational Sciences Division, USC, Seminar
| SCOUT - PAST PROJECT |
|---|
Project DescriptionSCOUT is an 2X2 MIMO-OFDM baseline system built on Sundance's SMT310Q mother board, which contains a Xilinx Virtex-II Pro FPGA board, and two SMT370 ADC/DAC boards. For the purpose of increasing the throughput, the system adopted a pipelined architecture using a Fast Fourier Transform that is shared across modulations of the system. The core of the baseline system is firstly modeled and simulated using Matlab Simulink. The prototype is then implemented using Xilinx System Generator for DSP. In order to make FPGA board communicates with the home PC via PCI interface, an I/O control unit plus asynchronous FIFOs acting as I/O buffers were designed and inserted as a wrapper to the original system. At last, to achieve flexibility with regard to data rates and effective communication range, the role of dynamic reconfiguration in the MIMO-OFDM systems analyzed and identified.
Affiliation
Dept of Electrical Engineering, USC. Sundance Multiprocessor Technology. Irvine Sensors Corporation, CA.
Research Projects @ PKU
| PHY layer All-digital PPM transceiver (ADPT) design - PAST PROJECT |
|---|
Project DescriptionADPT is an all-digital physical layer transceiver based on Pulse Position Modulation (PPM) scheme used in the field of wireless optical communication. The receiver consists of a mBnB (line code) Codec, a PPM Codec, and a second-order all-digital phase-locked-loop (ADPLL) for clock recovery, which is reconfigurable based on different precision requirements. The prototype transceiver was written in Verilog and built on Xilinx Virtex-E FPGA with an effective transmission rate of 10Mbps. An improved mBnB codec architecture was implemented to replace the original ROM-based architecture to save hardware resources and reduce the delay. Besides, an improved algorithm was also proposed and implemented to boost the speed of clock synchronization process for ADPLL. At last, a revised transceiver architecture based on Multiple PPM (MPPM) scheme was advanced for the purpose of improving the bandwith utilization.
Affiliation
Wireless Optical Communication National Laboratory, Dept of Electrical Engineering, School of Electronics Engineering and Computer Science, Peking University, Beijing, China.
Thesis (B.S.)
- Mi Zhang, "A PHY layer All-digital PPM transceiver (ADPT) on Xilinx Virtex FPGA" (in Chinese)  
