Endpoint Secured Multisensor Interface Chip for IoT Applications

The society cannot spin without data, including the vast amount of data collected by physical sensors. In the coming era of Internet of Things (IoT), the demand for sensor data is expected to grow exponentially with the emerging of advanced data analytics tools. Typically, IoT sensors collect physical data such as temperature, illuminance, electrical usage, air quality, water quality, or a patient’s vital signs, among others, and transmit it wirelessly to the network and/or management systems. These data can provide insights for a system to react timely on environmental changes. For example, an IoT sensor network can generate a detailed thermal and airflow map of a building, which helps to identify hotspots of the building and schedule energy-efficient cooling strategies. The broad use of IoT sensors could enable a smarter, healthier and low-carbon society. Usually, an IoT node includes multiple sensors to collect data from different physical sources, which requires the IoT node to convert different types of sensor signals accurately to their binary forms for processing and transmission. Meanwhile, in many use cases, the integrity and privacy of the IoT sensor data are safety-critical. It requires the IoT sensor to be resilient to both physical tampering and cyberattacks. It seems easy to satisfy these requirements by integrating different chip or hardware modules (e.g., data converters, MCU, cryptographic module, power management, etc.) on a single board. Unfortunately, IoT sensors have many unique features, including the ease of physical access, limited processing power and storage, cost-sensitive, and maintenance-free. Most solutions available in the market using discrete components are costly and power-hungry because they are not optimized at the system level. In addition, these solutions are unsecure as the available energy budget (e.g., sub-mW) of IoT sensors cannot support most of the existing crypto modules. Moreover, they are prone to physical tampering (e.g., replace a critical sensor with a malicious one) because of the lack of intelligent tools to detect the device abnormalities after deployment, except using frequent manual inspection. These factors are all impeding the large-scale deployment of IoT sensors for now. Many companies (e.g., TI, Maxim, NXP) are competing for the IoT sensor market. However, they focus more on the application-level integration and function enrichment instead of the front-end device optimization. Up to now, no widely adopted low-power IoT sensors with high data collection efficiency and security level are available in the market. To address this need and the above-stated technical issues, we will propose an integrated solution that is optimized at the chip level by leveraging the expertise of an interdisciplinary team and the supports from the end-users. This project aims to design a chip with a multisensor interface and an on-chip crypto engine for low-power IoT sensor network applications. The main features of this chip are: 1) it has an energy-efficient multisensor interface that can precisely convert different types of sensor signals to their digital forms, avoiding the use of additional discrete signal conditional chips. 2) it has a new scheme that can detect physical tampering with the sensor connections to avoid analog-domain attacks. 3) it has an energy-efficient NIST FIPS 140-3 compatible crypto engine for sensor node authentication and sensor data encryption, which is optimized at the algorithmic, hardware, and application levels. 4) it enjoys the benefits of hardware co-design of the sensing module and crypto engine for energy-efficient operation. The final integrated system will be in a SiP package with a small footprint (6mm x 6mm QFN). The developed chip will be firstly tested in-house for performance (in HBKU) and robustness verifications (in Silergy Corp.) and then be tested in an IoT sensor network demonstration system in Qatar (in QMIC). Potential technology transfer can be performed by Silergy Corp. once all the scheduled tasks are verified. Other outcomes of this project include the developed IPs and four well-trained researchers in the area of IC design and hardware crypto. It can potentially help Qatar to increase the research capacity and attract international investment in this area.

Hamad Bin Khalifa University (HBKU)