Information Agents for Autonomous Acquisition of Sensor Network Data A. Rogers and N. R. Jennings School of Electronics and Computer Science University of Southampton Southampton, SO17 1BJ, UK {acr,nrj}@ecs.soton.ac.uk M. A. Osborne and S. J. Roberts Department of Engineering Science University of Oxford Oxford, OX1 3PJ, UK {mosb,sjrob}@robots.ox.ac.uk ABSTRACT In this paper, we describe an information agent that can autonomously acquire sensor readings from environmental sensor networks (de- ciding when and which sensor to acquire readings from at any time). Moreover, this agent can perform a range of information processing tasks including modelling the accuracy of the sensor readings, predicting the value of missing sensor readings, and pre- dicting how the monitored environmental parameters will evolve into the future. We describe how our agent uses an iterative formu- lation of a multi-output Gaussian process to build a probabilistic model of the environmental parameters being measured by local sensors, and the correlations and delays that exist between them. We validate our approach using data collected from a network of weather sensors located on the south coast of England. Categories and Subject Descriptors I.2.11 [Computing Methodologies]: Artificial Intelligence—Dis- tributed Artificial Intelligence General Terms Algorithms, Experimentation Keywords information agent, sensor networks, Gaussian processes 1. INTRODUCTION Sensor networks have recently generated a great deal of research interest within the computer and physical sciences. Their use for the scientific monitoring of remote and hostile environments is in- creasingly common-place, and recent research has addressed how the information from such sensor networks can be made available to multiple users directly through standard web interfaces (see [4] for a review of such environmental sensor networks). Such sys- tems pose a number of novel challenges, not least the need for self- describing data formats, and standard protocols such that sensors can advertise their existence and capabilities to potential users of the network. However, more significantly for us, many of the information pro- cessing tasks that would previously have been performed by the owner or single user of an environmental sensor network (such as detecting faulty sensors, fusing noisy measurements from several sensors, and deciding how frequently readings should be taken) are now delegated to the multiple different users of the system, all of whom may have different goals and may be using sensor readings for very different tasks. Furthermore, the open nature of the net- work (in which additional sensors may be deployed at any time, and existing sensors may be removed, repositioned or updated) means that these users may have only limited knowledge of the precise location, capabilities, reliability, and accuracy of each sensor. Thus, there is a clear need for information agents that are capable of autonomously performing the acquisition and processing of in- formation from such sensor networks. Given this, in this paper, we describe our work developing just such an agent. This agent uses a novel iterative formulation of a multi-output Gaussian process (described in more detail in [6]) to build a probabilistic model of the environmental parameters being measured by local sensors, and then uses this model to perform a number of information process- ing tasks including: modelling the accuracy of the sensor readings, predicting the value of missing sensor readings, predicting how the monitored environmental parameters will evolve in the near future, and performing active sampling by deciding when and which sen- sor to acquire readings from. We use a network of weather sensors on the south coast of England to validate this approach, and we illustrate its effectiveness by benchmarking against the more con- ventional single-output Gaussian processes that models each sensor independently. 2. INFORMATION PROCESSING As discussed above, we require that our information agent be able to autonomously perform data acquisition and information process- ing despite having only limited specific knowledge of each sensor (e.g. their precise location, reliability, and accuracy). To this end, we require that it explicitly represent: 1. The noise in the sensor readings, and hence, the uncertainty in the environmental parameter being measured; sensor read- ings will always include measurement noise, and thus there will always be uncertainty in the agent’s world picture. 2. The correlations or delays that exist between sensor readings; sensors that are close to one another, or in similar environ- ments, will tend to make similar readings, while many phys- ical processes involving moving fields (such as the move- ment of weather fronts) will induce delays and correlations between sensors. We then require that the information agent use this explicit repre- sentation in order to perform: 1. Efficient active sampling by selecting when to take a reading, and which sensor to read from, such that the minimum num- ber of sensor readings are used to maintain an agent’s world uncertainty below a specified threshold (or minimising un- certainty given a constrained number of sensor readings).