Optimal detection of faulty traffic sensors used in route planning

Abstract

In a smart city, real-time tra c sensors may be deployed for various applications, such as route planning. Unfortunately, sensors are prone to failures, which result in erroneous tra c data. Erroneous data can adversely a ect applications such as route planning, and can cause increased travel time. To minimize the impact of sensor failures, we must detect them promptly and accurately. However, typical detection algorithms may lead to a large number of false positives (i.e., false alarms) and false negatives (i.e., missed detections), which can result in suboptimal route planning. In this paper, we devise an e ective detector for identifying faulty tra c sensors using a prediction model based on Gaussian Processes. Further, we present an approach for computing the optimal parameters of the detector which minimize losses due to false-positive and falsenegative errors. We also characterize critical sensors, whose failure can have high impact on the route planning application. Finally, we implement our method and evaluate it numerically using a realworld dataset and the route planning platform OpenTripPlanner. CCS CONCEPTS•Computer systems organization → Embedded and cyberphysical systems; Dependable and fault-tolerant systems and networks; • eory of computation → Gaussian processes; KEYWORDS fault detection, cyber-physical systems, smart city, route planning ACM Reference format:

Application-Aware Anomaly Detection of Sensor Measurements in Cyber-Physical Systems

“…We obtain the performance metrics and . Please note that more complex regression algorithms (e.g., Gaussian Processes [ 24 ]) can be used as well; however, we obtained satisfactory result with a simple linear regression model.…”

“…Please note that this approach is particularly applicable to traffic networks as there are usually many redundant sensors in the network. The function can then be obtained using suitable machine learning regression algorithm such as deep neural networks [ 23 ], Gaussian Processes [ 24 ], and many others [ 25 ]. Thus, for sensor s , we obtain the prediction as .…”

Machine Learning for Security and the Internet of Things: The Good, the Bad, and the Ugly

“…Nonetheless, the complexity of the network structure and its management pose high risks for data integrity. To combat this, Ghafouri et al [150] proposed an effective detector for inspecting sensor and communications failures to ensure data integrity in the smart city. The detector utilizes the Gaussian processes for optimization, as well as an approach for computing optimal parameters.…”

“…Evaluating the proposed detector on the OpenTripPlanner platform, the results showed that the detector could reliably increase data integrity in the smart city. Furthermore, Ghafouri et al [151] proposed a general framework that considered attacks on a subset of sensors in CPS, with specific emphasis on overcoming the limitations of their prior work [150]. This framework consists of a general anomaly detection module that predicts a measurement for each sensor and leverages three different regression models: linear, neural network, and combined.…”

Data-Driven Detection of Anomalies and Cascading Failures in Traffic Networks

“…The underlying assumption in this process is that the prior distribution of the regression function is considered to be a multivariate Gaussian distribution. By calculating the covariance matrix for the labeled data and covariance vector between labeled and new test data points and taking the measurement noise into account, the prediction result for the test data points can be obtained [Ghafouri, Laszka, Dubey, and Koutsoukos (2017)]. In this work, we have used Radial Basis Function (RBF) as the kernel.…”