[MWC 2024] AI stops before kickboard collision… IoT is also ‘on-device’

[MWC 2024] AI stops before kickboard collision…  IoT is also ‘on-device’

2024-02-27 23:00:00

Interview with Myung-Hoon Jeong, KT IoT Business Manager

“First year of on-device AIoT business”

IoT itself, such as sensors, performs AI calculations

Analyze accident video by attaching it to a kickboard

“By adding on-device artificial intelligence (AI) to the electric kickboard, it has become possible to stop the kickboard before a collision occurs.”

Myung-Hoon Jeong (picture) Managing Director of KT’s Digital Transformation (DX) Business Division’s Internet of Things (IoT) business is with reporters at the exhibition hall in Fira Gran Via, Barcelona, ​​Spain, on the 27th (local time), the second day of the world’s largest mobile communication exhibition ‘Mobile World Congress (MWC) 2024’. met with the company and introduced the company’s ‘On-Device Intelligent Internet of Things (AIoT)’ development status and announced that it would push forward with the business in earnest this year.

While on-device AI is an AI technology that embeds AI models directly into devices such as smartphones and can quickly calculate them without external connection, on-device AIoT combines this with IoT. It is a technology that allows IoT devices such as intelligent closed-circuit (CC)TV and sensors to directly perform AI calculations. For example, intelligent CCTV requires an external server through the cloud when performing AI calculations such as accident detection on recorded screens, but cameras made using the OndaVis AIoT method can perform the same task on their own.

Managing Director Jeong explained, “Currently, KT has secured three representative on-device AIoT technologies, including ‘AI electric vehicle charger,’ ‘AI scooter,’ and ‘smart signage.’” AI electric vehicle charger is a technology that detects fire and smoke when charging an electric vehicle through a camera attached to the charging station and quickly sounds an alarm. Not only does it prevent fire accidents during charging, which is a chronic problem with electric vehicles, but it also improves charging station management efficiency by detecting situations in which vehicles enter and leave the charging station and situations where the charger connector is left unattended.

KT joined hands with G-Bike and installed on-device AIoT technology in the form of a camera on electric kickboards. Managing Director Jeong said, “When a collision occurs while driving, AI motion sensing results can be uploaded to the server to help with post-accident processing, such as calculating insurance premiums.” He added, “The camera also measures the distance to surrounding objects and the environment.” “We can prevent a collision by recognizing it and slowing down or stopping if it gets too close,” he explained. He added, “We are also considering a business model that provides compensation to users who park their kickboards in six suitable parking spaces, such as kickboard parking lots and next to street trees, and warns users who do not.”

For the third time, KT also installed on-device AIoT on signage, which is an advertisement board above taxis. Basically, by analyzing the floating population and surrounding commercial areas through cameras, customized advertising is possible for each area through which the taxi passes. Furthermore, it is equipped with sensors that can measure location, weather, traffic, harmful gases, fine dust, etc., and collects data in various fields such as smart cities in addition to advertising purposes, and trades it or provides it to the public such as local governments. The model is also being reviewed. Managing Director Jeong said, “We joined hands with Motov to introduce smart signage to 2,500 taxis in Seoul. We can collect (desired) data in one hour.”

He said, “We will use this year as the first year of the on-device AIoT business,” and added, “We plan to expand the application of on-device AIoT equipment not only to electric vehicle charging stations, but also to places such as hiking trails where CCTV is difficult to install, crime blind spots, and remote remote areas.” .

< 저작권자 ⓒ 서울경제, 무단 전재 및 재배포 금지 >

1709076178
#MWC #stops #kickboard #collision.. #IoT #ondevice

Leave a Replay