Chien-Liang Fok

Challenges of satisfying multiple stakeholders: quality of service in the internet of things

As wireless sensor networks become increasingly integrated with the Internet, the Internet of Things (IOT) is beginning to emerge. The IOT is a massive cyber-physical system that presents many software engineering challenges. One fundamental challenge is the need for multi-dimensional QoS that can satisfy the individual constraints of the many participants in the system. In this paper, we investigate the challenges of providing such a mechanism via a simple abstraction consisting of a general QoS function provided by each application. This function distills the multi-dimensional QoS specifications from each stakeholder into a single value that is used to determine the best configuration of interactions. We prototype our approach in a real wireless sensor network using a pervasive healthcare fall-detection application and highlight the many challenges it unveils.

A Platform for Evaluating Autonomous Intersection Management Policies

There is a significant push towards greater vehicular autonomy on roads to increase convenience and improve overall driver experience. To enable this autonomy, it is imperative that cyber-physical infrastructure be deployed to enable efficient control and communication. An essential component of such road instrumentation is intersection management. This paper develops an intersection management platform that provides the sensing and communication infrastructure needed to enable efficient intersection management policies. The test bed, located in a indoor laboratory, consists of an intersection and multiple robotic vehicles that can sense and communicate. Whereas traditional approaches to intersection management rely on simulations, this test bed enables the first realistic evaluation of several intersection management policies. Six simple but practical centralized and distributed policies are evaluated and compared against the current state of the art, i.e., traffic signals and stop signs. Through extensive experimentation, this paper concludes that, in the scenario tested, even a simple coordinated management policy can halve vehicular delay, while improving the aggregate traversal time of the intersection by 169%.

Evasion planning for autonomous vehicles at intersections

Autonomous intersection management (AIM) is a new intersection control protocol that exploits the capabilities of autonomous vehicles to control traffic at intersections in a way better than traffic signals and stop signs. A key assumption of this protocol is that vehicles can always follow their trajectories. But mechanical failures can occur in real life, causing vehicles to deviate from their trajectories. A previous approach for handling mechanical failure was to prevent vehicles from entering the intersection after the failure. However, this approach cannot prevent collisions among vehicles already in the intersection or too close to stop because (1) the lack of coordination among vehicles can cause collisions during the execution of evasive actions; and (2) the intersection may not have enough room for evasive actions. In this paper, we propose a preemptive approach that pre-computes evasion plans for several common types of mechanical failures before vehicles enter an intersection. This preemptive approach is necessary because there are situations in which vehicles cannot evade without pre-allocation of space for evasion. We present a modified AIM protocol and demonstrate the effectiveness of evasion plan execution on a miniature autonomous intersection testbed.

On coordination in practical multi-robot patrol

Multi-robot patrol is a fundamental application of multi-robot systems. While much theoretical work exists providing an understanding of the optimal patrol strategy for teams of coordinated homogeneous robots, little work exists on building and evaluating the performance of such systems for real. In this paper, we evaluate the performance of multirobot patrol in a practical outdoor distributed robotic system, and evaluate the effect of different coordination schemes on the performance of the robotic team. The multi-robot patrol algorithms evaluated vary in the level of robot coordination: no coordination, loose coordination, and tight coordination. In addition, we evaluate versions of these algorithms that distribute state information-either individual state, or entire team state (global-view state). Our experiments show that while tight coordination is theoretically optimal, it is not practical in practice. Instead, uncoordinated patrol performs best in terms of average waypoint visitation frequency, though loosely coordinated patrol that shares only individual state performed best in terms of worst-case frequency. Both are significantly better than a loosely coordinated algorithm based on sharing global-view state. We respond to this discrepancy between theory and practice, caused primarily by robot heterogeneity, by extending the theory to account for such heterogeneity, and find that the new theory accounts for the empirical results.

Network coded routing in delay tolerant networks: an experience report

In delay-tolerant networks, end-to-end routes are rarely available, and routing protocols must take advantage of the opportunistic interactions among nodes to deliver packets. Probabilistic routing performs well in such networks and has been the dominant focus of research in this area. However, creating efficient routing protocols is challenging because to reduce latency, one often needs to replicate messages thus increasing routing overhead. Network coding has been explored as a way to increase throughput in DTNs without a significant increase in overhead, and network coded routing approaches have shown promising results. In this paper, we report on our experience integrating both erasure coded and network coded routing into the well-adopted DTN2 Reference Implementation. We implement our routing module and evaluate it via small real-world field tests.

Passive network-awareness for dynamic resource-constrained networks

As computing becomes increasingly mobile, the demand for information about the environment, or context, becomes of significant importance. Applications must adapt to the changing environment, however acquiring the necessary context information can be very expensive because collecting it usually requires communication among devices. We explore collecting reasonably accurate context information passively by defining passively sensed context through network overhearing, defining context metrics without added communication cost. We use this framework to build a small suite of commonly used context metrics and evaluate the quality with which they can reflect ground truth using simulation. We also provide an implementation of this passive sensing framework for Linux, which we evaluate on a real mobile ad-hoc network using mobile autonomous robots with commodity 802.11 b/g wireless cards.

Brace: an assertion framework for debugging cyber-physical systems

Developing cyber-physical systems (CPS) is challenging because correctness depends on both logical and physical states, which are collectively difficult to observe. The developer often need to repeatedly rerun the system while observing its behavior and tweak the hardware and software until it meets minimum requirements. This process is tedious, error-prone, and lacks rigor. To address this, we propose BRACE, A framework that simplifies the process by enabling developers to correlate cyber (i.e., logical) and physical properties of the system via assertions. This paper presents our initial investigation into the requirements and semantics of such assertions, which we call CPS assertions. We discusses our experience implementing and using the framework with a mobile robot, and highlight key future research challenges.

MobiQuery: A Spatiotemporal Query Service for Mobile Users in Sensor Networks

This paper presents MobiQuery, a spatiotemporal query service that allows mobile users to periodically collect sensor data from the physical environment through wireless sensor networks. A salient feature of \MQ is that it can meet stringent spatiotemporal performance constraints, including query latency, data freshness, and changing areas of interest due to user mobility. We present three just-in-time prefetching protocols that enable MobiQuery to achieve desired spatiotemporal performance despite low node duty cycles, while significantly reducing communication overhead. We validate our approach through both theoretical analysis and extensive simulations under realistic settings including varying user movement patterns and location errors.

A Unified Specification Framework for Spatiotemporal Communication

Traditionally, network communication entailed the delivery of messages to specific network addresses. As computers acquired multimedia capabilities, new applications such as video broadcasting dictated the need for real-time quality of service guarantees and delivery to multiple recipients. In light of this, a subtle transition took place as a subset of IP addresses evolved into a group-naming scheme and best-effort delivery became subjugated to temporal constraints. With recent developments in mobile and sensor networks new applications are being considered in which physical locations and even temporal coordinates play a role in identifying the set of desired recipients. Other applications involved in the delivery of spatiotemporal services are pointing to increasingly sophisticated ways in which the name, time, and space dimensions can be engaged in specifying the recipients of a given message. In this paper we explore the extent to which these and other techniques for implicit and explicit specification of the recipient list can be brought under a single unified framework. The proposed framework is shown to be expressive enough so as to offer precise specifications for existing communication mechanisms. More importantly, its analysis suggests novel forms of communication relevant to the emerging areas of spatiotemporal service provision in sensor and mobile networks. Currently at Palo Alto Research Center, 3333 Coyote Hill Road, Palo Alto, CA 94040

A Lightweight Coordination Model and Middleware for Mobile Computing **Please see WUCSE-03-12**

LimeLite is a new coordination model and middleware designed to support rapid development of applications entailing logical mobility of agents and physical mobility of hosts. Designed to function in open environments, LimeLite performs automatic agent discovery but filters the results to define for each agent an individualized acquaintance list in accordance with run-time policies specified at the application level. This asymmetry among participants in the coordination process is dictated by the need to accommodate settings involving large numbers of agents and hosts that come and go freely. It represents an important departure from coordination research in general. The coordination context is limited to the specific needs of the individual agent and its coordination activities are restricted to tuple spaces owned by peers present in the acquaintance list. Linda-like primitives typically used in coordination middleware are tailored in LimeLite to address the challenges of mobile environments. Among other things, this entails the elimination of remote blocking and data pushing operations since the affected agents may no longer be within communication range. It also entails the addition of reactions that are triggered by the presence of information of interest on agents listed in the acquaintance list and not by events that could have occurred prior to discovery. Finally, to ensure both performance and ease of deployment on small devices the granularity of atomic operations and the reliance on transport layer guarantees have been minimized. This paper introduces LimeLite, explains its key features, illustrates its usage in application development, and explores its effectiveness as a software engineering tool.