Real-Time Systems and Sensor Networks Sang H. Son Department of Computer Science University of Virginia Charlottesville, Virginia 22904 University of Virginia Input Real Real-Time (Embedded) System World Output • Input – current state (view) update – tasks to be performed by real-time systems • Output – actions to change real world situation – information to be used to support decision-making University of Virginia Real-Time Systems • Real-time systems – typically embedded in a large complex system – timeliness and dependability (reliability) are crucial – explicit/implicit timing constraints (soft, firm, hard) • A large number of applications – aerospace and defense systems, nuclear systems, robotics, process control, agile manufacturing, stock trading, network and traffic monitoring and control, multimedia computing, databases, medical systems, wireless sensor networks • Rapid growth in research and development – workshops, symposia, journals – standards (RT-Linux, RT-Java, RT-COBRA, …) University of Virginia Time Constraints v(t) v0 d t v(t) v0 d1 d2 t University of Virginia Trends in Real-Time Systems Applications • Soft real-time requirements rather than hard ones – much wider applications – relates well with the notion of QoS – soft is harder to deal with than hard ones • Operate in unpredictable environments – WCET too pessimistic or high variance – unbounded arrival rate; overload unavoidable • Need to support multi-dimensional requirements – real-time, power, size, security, and fault-tolerance – conflicting resource requirements and system architecture • Embedded and interacting with physical world University of Virginia Key Issues (Part of a Long List) • Real-time services in embedded networked systems – flexible and adaptable (self-configurable) – interaction with physical/distributed environment sensors/actuators in mobile nodes using WSN – group-based aggregation and confidence management – scalability • Multi-dimensional constraints – real-time, location-dependence, power, mobility, wireless, size, cost, fault-tolerance, security and privacy • Timely management of real-time data (QoD/QoS) – large volume with temporal properties – robust real-time data and event services University of Virginia QoS Management in Real-Time Data • Motivation – increasing demands for real-time data/event services • web-based information services and e-business • sensor networks • interactive rendering • location-aware services in mobile networks – temporary overload and service degradation inevitable • Service quality: QoS parameters – timeliness – data freshness – degree of imprecision – behavior in transient state: overshoot and settling time University of Virginia Feedback Control control input Controller feedback Actuator Process Sensor controlled variable reference (set point) University of Virginia Timeliness Specification Miss ratio Overshoot Steady state error % Reference Transient State Steady State Time Settling time University of Virginia Data Freshness Database Database Freshness: Set of continuous data Perceived Freshness: Set of continuous data accessed by timely transactions University of Virginia WSN Application Spectrum Interactive VR Game Environmental Monitoring Wearable Computing Disaster Recovery Earth Science & Exploration Immerse Environments Wireless Sensor Networks Hazard Detection Military Surveillance Context-Aware Computing Biological Monitoring Smart Environment Linear Structure Protection Urban Warfare University of Virginia Data/Event Services in Sensor Networks • Recent advances in low-cost low-power devices – large scale sensor networks (ad hoc mobile networks) – each node consists of sensors/actuators/processors • Issues in wireless sensor networks – how to collect and disseminate real-time data – QoS management under resource constraints – how to conserve energy while satisfying application requirements – efficient real-time localization – consensus, aggregation, in-network processing, confidence, security University of Virginia Event Services for Emergency Response Technology and Research • Confidence levels in data • Multi-level events • Real-Time • Minimize false alarms • Actual implementation Multidisciplinary • Sensor design • Application to emergency response services Impact • Save lives • Minimize damage • Improve response to natural disasters or terrorist attacks Dynamic Deployment of Wireless Sensor Network (self-organizing) Evacuate people ahead of leak Explosion Gas Leak Simple event reports Detect compound event & dispatch emergency rescue team University of Virginia Undersea Surveillance • Automated real-time undersea surveillance project by Navy • Acoustic communication in undersea • Experiments performed in 2003 in Florida • Three sensor nodes in a cluster – Each had 3-dimensional magnetic sensor • One submarine at a time moved through the network • Data was gathered during experiments and analyzed later (not real-time) University of Virginia Issues in Undersea Surveillance • Feature extraction from magnetic/acoustic sensors to mitigate false alarms • Confidence to reduce false positives/negatives • System architecture and trade-off analysis • Identify system objectives and key performance parameters – System configuration (# and type of sensor nodes, surveillance coverage, deployment, …) – System parameters (sensing/communication ranges, duty cycles, data aggregation ratio, …) – Adaptation under uncertainty University of Virginia Undergrad Research Assistant • NSF REU (Research Experience for Undergrads) • What is promised – challenging problems – rewarding research experience • Offers will not last long – call toll-free number 1-800-982-2205 – e-mail to son@cs.virginia.edu University of Virginia