SMACS Research Compilers Programming Language Database Sensor Networks Computer Vision Machine Learning Natural Language Processing Bioinformatics Robotics High Performance Computing Pervasive Computing
Faculty Research Interests See individual web pages for complete details
David Hsu, Tomas Lozano-Perez
Compilers and Programming Languages
Chin Wei Ngan, Wong Weng Fai, Martin Rinard
Ooi Beng Chin, Tan Kian Lee, Stuart Madnick
Lee Wee Sun, Leslie Kaelbling, Tomas Lozano-Perez
Natural Language Processing
Ng Hwee Tou, Lee Wee Sun
David Hsu, Leslie Kaelbling, Tomas Lozano-Perez
Pervasive Computing, Computer Vision
Cham Tat Jen, Larry Rudolph
Parallel and Distributed Computing
Hsu Wen Jing, Teo Yong Meng, Charles Leiserson, Alan Edelman
Current PhD Students Lee Wee Sun, Leslie Kaelbling Multisignal classification and learning Chieu Hai Leong David Hsu, Tomas Lozano-Perez, Leslie Kaelbling An Intelligent Robot Tracker for Elderly Care Amit Jain David Hsu, Tomas Lozano-Perez, Leslie Kaelbling An Intelligent Robot Tracker for Elderly Care Jiang Xiaoxi Advisors Project Student
Hsu Wen Jing Adaptive Resource Discovery in Distributed Networks Fang Hui Ooi Beng Chin Semantics Based Information Retrieval in P2P based System Mihai Lupu Ooi Beng Chin Query Processing in Structured P2P Network Vu Quang Hieu Tan Kian Lee Mobile Peer-based Data Management Wu Wei Leong Tze Yun System Biology Zhu Ailing
Ooi Beng Chin, Karen Sollins P2P data/content sharing systems and data integration Yu Bei Wong Weng Fai, Larry Rudolph Dynamic Code Optimization Zhao Qin Leong Tze Yun, Leslie Kaelbling Reinforcement Learning Ong Chen Hui Chin Wei Ngan, Martin Rinard Sized Region for Real-Time Java Nguyen Huu Hai Hsu Wen Jing, Charles Leiserson Adaptive Distributed Services Based on Peers Technology He Yuxiong David Hsu, Tomas Lozano-Perez Algorithms for Understanding Protein Structural Flexbility Anshul Nigam
Computers will be everywhere embedded into the environment Computation will be virtually free But how do we exploit that ....
Motes and smart dust ...
wireless sensor platform to provide the flexibility to create powerful, wireless, and automated data collection and monitoring systems
hardware platform consists of Processor/Radio boards (MPR) commonly referred to as MOTES
battery-powered devices run TinyOS and support two-way mesh radio networks
Senses temperature, ambient light, vibration, acceleration, or air pressure, etc., processes the data, and stores it in memory
Vision: In 2010 MEMS sensors will be everywhere, and sensing virtually everything. Scavenging power from sunlight, vibration, thermal gradients, and background RF, sensors motes will be immortal, completely self contained, single chip computers with sensing, communication, and power supply built in. Kris Pister, UC Berkeley I 2 R will be building wearable devices based on these types of devices ...
GPS outdoors ... Crickets indoors...
Distance ranging and positioning precision of between 1 and 3 cm
Designed for low-power operation
Can be used as a location-aware sensor computing node (running TinyOS), to which a variety of sensors can be attached.
Raw data, such as: Accelerometer Heat flux Galvanic skin response Skin temperature Event timestamp
Derived data, such as: Total calories burned Duration of physical activity Number of steps Resting energy expenditure Active energy expenditure Sleep onset Wake time Sleep duration
Contextual data, such as: Ambulatory exercise Lying down In/Out of bed Sleeping On/Off body
Mobile robots ....
Vision - able to capture, recognize and identify thousands of objects and locations
Hearing - contains 'listen for' speech recognition; can also respond to sound levels
Speech - able to talk using a 'phrase to speak' function
Networking - can send and receive e-mail; able to e-mail commands when used with a wireless card and network
Remote control - can be teleoperated from an external computer (networking equipment outlined above)
Autonomous mobility - able to specify movement parameters such as direction and target characteristics, allowing the robot to move around by itself
Gripping ** - optional arm-like Gripper grabs and carries objects
IR sensing ** - optional IR Sensor Pack provides object presence detection, allowing you to trigger a behavior if an object is detected