Novel Data Framework for Efficient Smart City Security Systems
2023 Best Paper Award Winner
Researchers develop a framework for efficiently processing the large data generated by multi-agent smart security systems
Smart cities will use multiple agents to collect environmental data and provide security and convenience to their residents. However, the processing of the large data generated in these multi-agent systems poses a challenge. To address this, researchers have developed a novel data framework that efficiently processes data from multiple agents to generate an observation map and uses an anomaly detection algorithm to detect abnormal activities. This method can reduce costs for realizing practical smart cities.
Smart cities, equipped with advanced technologies, are on the horizon. They can provide their residents with security and convenience by utilizing multiple agents like CCTV cameras, delivery robots, unmanned vehicles, and security robots. These agents play an important role in collecting environmental data that can then be used for various applications including advertising and security monitoring.
While many studies have attempted to combine these agents to provide security services to reduce costs and address blind spots, processing the large, complex, and dynamic data generated in these systems poses a significant challenge. Moreover, due to the current limitations of wireless communications, data processing is performed within the agents themselves, which further complicates the issue.
In response to these problems, a team of researchers from Korea, led by Dr. Hochul Shin from the Department of Intelligent Robotics at the Electronics and Telecommunications Research Institute, has now developed a novel framework for efficiently processing large data. “In the future, activities such as delivery robots, security robots, and unmanned shuttles will increase in people’s daily living environments. Using our method, the data on various situations and events that occur in these situations can be effectively collected, stored, and utilized,” says Dr. Shin. Their study was published in Volume 44, Issue 2 of the ETRI Journal on April 25, 2022.
The innovative algorithm uses various mobile and fixed robots in a city to collect information about their environments. Dr. Shin draws parallels to human memory, stating: “Just as people remember previous situations about places they frequently visit and compare them with the current situation, a monitoring system linked to fixed and mobile robots, monitors various events and situations that occur in areas designated by the user.” This data is then used to create a local surveillance map of a specific area, which, in turn, is sent to a central data server where multiple local maps are combined to form a global observation map. Alongside detecting pedestrians and vehicles, this map can also detect changes in elevation and temperature.
Moreover, the researchers also developed an anomaly detection system that utilizes this surveillance map data to detect abnormal situations. They tested their framework in a designated area with multiple fixed and mobile agents for four months. Their experiment revealed that their algorithm effectively detects movements of pedestrians, cars, and anomalies such as unusual crowds. It also successfully identified elevation changes from falling trees and container movement as well as temperature changes from welding and vehicles as abnormal. Importantly, the algorithm could effectively and efficiently process the data generated by multiple agents in real-time.
This groundbreaking algorithm has the potential to improve security systems in smart cities while simultaneously reducing costs through efficient processing of large data. Moreover, its applications extend to various other scenarios where environmental data is utilized by multiple agents in cities.
Overall, this study paves the way for efficient multi-modal systems, bringing us closer to realizing practical smart cities!
Reference
Title of original paper: Multimodal layer surveillance map based on anomaly detection using multi-agents for smart city security
Journal: ETRI Journal
About the Electronics and Telecommunications Research Institute
Established in 1976, ETRI is a non-profit government-funded research institute and is one of the leading research institutes in the wireless communications domain. It has more than 2500 patents filed. Equipped with state-of-the-art labs, this institute strives for social and economic development through technology research.
About the author
Hochul Shin is currently a senior researcher at the Electronics and Telecommunications Research Institute, Daejeon, Republic of Korea. His research focuses on using robotics technology to improve human life. Since receiving his Ph.D. from KAIST in 2005, he has been conducting research and development in fields such as surgical robots, service robots, and security robots. Recently, he has been researching wearable robots to improve the lives of workers, older adults, and people with physical weakness.
