Research

Overview of Our Research

We conduct a wide range of research to improve the safety and security of embedded systems. Our work is based on fundamental technologies such as scheduling theory for embedded systems, operating system (OS) design, implementation, and verification. Using these foundations, we study system analysis methods, partitioning techniques, and testing techniques (especially fuzzing).

We actively participate in joint research with industry and industry–academia–government collaboration projects. Through these activities, we have worked with companies and organizations in various fields, including automotive, railway, and aerospace systems.

In the past, many embedded systems were relatively simple, with small-scale software. This made it easier for designers to understand the entire system and to ensure not only performance, but also overall reliability, including safety and security (often called dependability).

However, modern embedded systems have become much more large-scale and complex, both in hardware and software. In addition, multiple devices are now connected through networks to provide new services.

In our research, we take two complementary perspectives.

One is the traditional embedded-system perspective, where we build systems using well-understood components and guarantee performance and quality step by step.

The other is the Systems-of-Systems (SoS) perspective, which focuses on large-scale systems that include unknown or continuously changing elements, as well as systems that interact with humans. From this viewpoint, our goal is to analyze and understand such systems and to explain their behavior while keeping them in the best possible state over time.

In other words, our research covers a broad range of topics—from deep technical research on system software (practical, hands-on research to improve software technologies) to future-oriented research that treats social and technical systems as Systems of Systems. In the latter, we extend ideas from economics and multi-agent systems to develop methods for modeling, designing, and verifying SoS, with an ideal future society in mind.

We welcome not only students from the Computer Science (CS) program, but also students from other departments and people from outside the university. If you are interested in discussing research topics or joining or collaborating with our research group, please feel free to contact us.

Real-Time Performance Assurance in High-Performance Embedded Real-Time Systems

In recent years, control computers equipped with high-performance System-on-Chip (SoC) platforms and operating systems have been widely used in automotive and aerospace systems. These platforms increasingly integrate and execute multiple applications with different levels of reliability and criticality at the same time. Such systems are called Mixed-Criticality Systems (MCSs).

When integrating multiple applications, it is essential to ensure real-time performance and safety. However, applications often share system resources—such as CPUs, main memory, GPUs, storage devices (eMMC, UFS, SSD), and network bandwidth—which can lead to problems such as fault propagation and security attacks. In addition, interference in access to shared resources can delay application execution. This may violate timing deadlines and degrade real-time guarantees, potentially causing serious harm to the system.

In this research, we measure and analyze access to shared resources by real-time–critical applications and evaluate their real-time performance to identify factors that degrade real-time behavior. Furthermore, we study a wide range of techniques to improve real-time performance, taking a full-stack approach, including the development of software-based isolation mechanisms, improvements to operating systems and device drivers, and the design of hardware mechanisms.

IoTxWeb3: A Framework for Automating Data Sharing and Lifecycle Management of IoT Devices

The number of network-connected embedded devices (IoT devices) is rapidly increasing. IoT devices range widely—from relatively inexpensive products such as smartwatches, smart locks, and robot vacuum cleaners, which are used by people of all ages in ordinary households, to more expensive systems such as those used in automobiles, aircraft, and industrial machines, which are operated and managed by trained engineers with specialized knowledge and licenses.

When considering the lifecycle of IoT devices—from design and manufacturing to operation and disposal—high-end IoT devices are typically managed by professionals. This makes it easier to ensure overall reliability, including safety, security, and fault handling, and also allows device management itself to become a viable business.

In contrast, consumer IoT devices are usually operated by general users, who rely on manuals and software updates (often via smartphones) to maintain safety and security when vulnerabilities are discovered.

At the same time, without users being fully aware—or with their consent implicitly included in long and complex terms of service—various types of data, such as device operation logs and sensor data, are often transmitted to device manufacturers via the Internet. These data are used for purposes such as device monitoring and AI training. Even if users want to make use of their own device data, they are often unable to do so because detailed device information and interfaces are not normaly available.

In this research, we develop a software framework that combines IoT devices with Web3 technologies, including blockchain and smart contracts, to enable automated lifecycle management and data sharing for IoT devices. Users can grant access to their IoT devices to manufacturers or third parties, and all permissions for device operation and data usage are automatically enforced through smart contracts on the blockchain. By recording all contracts on the blockchain, the framework ensures verifiability and transparency.

The IoTxWeb3 framework is being released as open-source software at our Github, allowing anyone to easily try and experiment with it.

Modeling, Analysis, and Dependability Assurance of Systems of Systems (SoS)

A System of Systems (SoS) refers to a structure in which multiple independent systems cooperate to provide services. Examples include road traffic systems, smart homes, and Mobility-as-a-Service (MaaS). More familiar examples can be found in everyday life, such as families or sports teams, where individual people act independently while cooperating to achieve a common goal. In fact, many systems in the real world can already be understood as Systems of Systems.

In SoS, the configuration and functions of individual systems, the connections between systems, and the quality of services are constantly changing. Because of this, it is difficult to anticipate and analyze all possible situations at the design stage and to build mechanisms that ensure safety and security in advance.

In this research, we model and analyze socio-technical systems as Systems of Systems and develop theories and technologies to ensure overall dependability, including safety, security, and resilience. Our goal is to engineer mechanisms—often referred to as SoS Engineering—that maintain and improve SoS dependability not only at the design stage but also through continuous analysis during system operation.

For SoS modeling, we apply and compare various modeling approaches, such as the business model description language ArchiMate, the system-theoretic accident model STAMP, and the Functional Resonance Analysis Method (FRAM). We also explore simpler representations based on graph structures.

In addition, we study game-theoretic consensus-building methods to help diverse stakeholders—such as users, designers, operators, maintenance engineers, and policy or rule makers with different backgrounds and expertise—reach agreement on failures, incidents, and response strategies. We also work on assurance and argumentation techniques using Goal Structuring Notation (GSN) to explain how dependability is achieved in Systems of Systems.

Research

Overview of Our Research

Real-Time Performance Assurance in High-Performance Embedded Real-Time Systems

IoTxWeb3: A Framework for Automating Data Sharing and Lifecycle Management of IoT Devices

Modeling, Analysis, and Dependability Assurance of Systems of Systems (SoS)

Other research topics and software発

Automotive Control Software Platforms

Embedded Real-Time Operating Systems

Software Platforms for IoT and Educational Systems

Tools and Simulators

In-Vehicle Network Simulation and Real-Time Analysis

Embedded Security