Anomaly Detection for Sensor Manipulation in Matter Enabled-IoT Devices with Faulty Data Injection
Jahnvi Patel, Kapil Sharma · 2025
The security of Internet of Things (IoT) devices in Matter-enabled networks is crucial, as attacks such as sensor spoofing and data injection can lead to significant disruptions. Sensor manipulation attacks where the attackers manipulate the sensor inputs to create erroneous behavior in devices are issues of serious concern for anomaly detections. Faulty sensor data can cause devices to operate in unsafe or inefficient conditions across various applications, compromising system reliability and performance. The Isolation Forest algorithm is employed to isolate faulty data injections as outliers in sensor readings, offering a robust method for anomaly detection. The approach, therefore, focuses on the possibility of enhancing detection with greater reliability in IoT networks. The lightweight and extensible approach also assembles a process of strengthening Matter-based networks against vulnerabilities emanating from wrong control feedback.