Hide menu

Research areas

Currently, most of my research is related to the following three topics.

Inertial sensor arrays

Today, thanks to development of the micro-electromechanical-systems (MEMS) technology, miniaturized accelerometers, gyroscopes, and magnetometers can be produced at unprecedented volumes and at low price. Unfortunately, the performance of these ultra-low-cost sensors is too poor to enable accuracte and reliable localization in many applications. However, by capitalizing on the small size, low price, and low power consumption of the sensors it is now feasible to construct arrays with hundreds of sensing elements and fuse their measurement to create "super sensors" with unprecedented performance-to-price ratios;
In-house developed smart inertial sensor array with 288 sensing elements. The range, sampling time, and bandwidth of the individual sensors can be controlled via the software of the embedded processor. This enables the array to adapt to the usage conditions.
However, before these "super sensors" can become real, several challanges related to: how to fuse the information from the sensors; how to calibrate the sensor arrays; and how to automatically adapt the sensors settings to different usage conditions; needs to be solved. I together my colleagues have looked at some of these challanges in the following papers:


Today more than 54% of the world population lives in urban areas - a figure that is continuously growing. The transport logistics problems arising from this urbanization call for intelligent transportation systems with increased efficiency, capacity, and reliability for not only horizontal but also vertical transportation. To adapt the vertical transportation systems to the demands of the future, the elevator industry has identified a need to move from today's preventative and corrective maintenance strategies to predictive and preemptive maintenance strategies. Accordingly, the latest generation of high-end elevator systems is frequently connected to the cloud, creating an Internet-of-Elevators, where data from the elevator sensors and control systems is gathered, mined, and transformed into information about the elevator systems' performance and any current or potential problems. However, the majority of today's elevator systems are not equipped with sensors and control systems that support connection to the cloud for remote monitoring and fault diagnostics, and it will take decades before they are all renewed or upgraded with such systems. In the interim, there will be a need for easy-to-install sensor nodes by which existing elevator systems can be upgraded and connected to the Internet-of-Elevators.

Therefore, I'm in collaboration with SafeLine Sweden AB, currently conducting research on signal processing methods that enables the creation of a smart sensor node that can be used for non-invasive condition and fault monitoring of elevators. A smart sensor node characterized by non-intrusive sensing via accelerometers and magnetometers; embedded data processing and storage; and wireless connectivity; enabling the construction of a truly "plug-and-play" node that can provide high-level elevator condition information. Some initial results from the research can be found in the paper:
Envisioned Internet-of-Elevator system for condition monitoring of elevators. Smart sensor nodes non-intrusively extract condition information using built-in accelerometer and magnetic field sensors. The condition information is sent to a cloud server where it is analyzed and combined with historical maintenance statistics to determine the maintenance strategy. The maintenance need is communicated to the service technician, who then takes appropriate actions and reports back to the cloud server the adjustments or repairs that have been made.

Insurance telematics

Insurance telematics is a disruptive technology that is expected to reform the vehicle insurance industry. Based on sensor data, the traditional measures for calculating the insurance premium are complemented to determine a fee that more accurately predicts the risk profile of the policyholder. From an signal processing point of view, there are several insurance telematics challenges that have to be tackled. It is about consistently, from low quality data, extracting relevant figures of merit (FoMs) like number of harsh braking, speeding, heavy cornering, trip smoothness, etc, and then to transform these FoMs into a valid measure, or score, that determines the risk profile of the insurance customer.

During the last 4 years I, together with Prof. P. Handel (KTH) and J. Wahlstrom (KTH), have conducted research on how the smartphone can be used as a measurement probe in insurance telematics applications. The main focus has been on developing signal processing methods that enables (reliable) calculation of driver behavior FoM from the often poor data provided by the smartphone sensors. Some of our most recent publications on the topic are:
Isaac Skog

Assistant Professor in Automatic Control,
Docent in Signal Processing

(Swedish: Biträdande universitetslektor i reglerteknik)

+46 708186805
Dept. of Electrical Engineering
Linköping University
SE-581 83 Linköping
Visiting Address:
Campus Valla
Building B
Room 2A:526

Page responsible: Isaac Skog
Last updated: 2017-09-19