SUM Tests Kia PV5 Autonomous Driving

SUM is validating autonomous driving technology on public roads with Kia’s PV5 purpose-built vehicle (PBV), which received a temporary operation permit. The company is expanding next-generation AI autonomous-driving technology to eco-friendly buses and PBVs, and plans to use PV5 operations to train AI and validate technology across different vehicles.

Founded in 2020, SUM previously collected autonomous-driving data from Hyundai’s Elec City and County Electric buses in different operating environments and for different purposes. It plans to add the Kia PV5 and gather vehicle-specific operating data from buses, PBVs and logistics vehicles, then use the data for AI training and technology validation required by each service. SUM says it is training AI with experience data gathered by real vehicles on roads to continually improve their ability to handle autonomous driving.

The PV5 autonomous-driving technology uses SMOBI-ToD, a remote-driving solution. The solution serves as a safety measure for situations that autonomous-driving systems have difficulty handling, and is also used to collect data on real road conditions and driver behavior. SUM plans to train AI on this data to commercialize next-generation end-to-end (E2E) autonomous-driving technology, which processes inputs through outputs as one integrated system. Existing technology handles perception, decision-making and control in separate, sequential systems; E2E uses road conditions and driver steering, acceleration and braking data in an AI model.

SUM plans to conduct a remote-driving demonstration with the PV5 at Pangyo Techno Valley in Gyeonggi Province this year and validate AI training results in a real vehicle. CEO Hyun Young-jin said that building real-road experience with different vehicles and improving technology with AI determine autonomous-driving competitiveness. The company said it aims to advance its technology, including obtaining a temporary operation permit in 2027 for a vehicle using E2E autonomous driving, and grow into a physical AI mobility company.

SUM is validating autonomous driving technology on public roads with Kia’s PV5 purpose-built vehicle (PBV), which received a temporary operation permit. The company is expanding next-generation AI autonomous-driving technology to eco-friendly buses and PBVs, and plans to use PV5 operations to train AI and validate technology across different vehicles.

Founded in 2020, SUM previously collected autonomous-driving data from Hyundai’s Elec City and County Electric buses in different operating environments and for different purposes. It plans to add the Kia PV5 and gather vehicle-specific operating data from buses, PBVs and logistics vehicles, then use the data for AI training and technology validation required by each service. SUM says it is training AI with experience data gathered by real vehicles on roads to continually improve their ability to handle autonomous driving.

The PV5 autonomous-driving technology uses SMOBI-ToD, a remote-driving solution. The solution serves as a safety measure for situations that autonomous-driving systems have difficulty handling, and is also used to collect data on real road conditions and driver behavior. SUM plans to train AI on this data to commercialize next-generation end-to-end (E2E) autonomous-driving technology, which processes inputs through outputs as one integrated system. Existing technology handles perception, decision-making and control in separate, sequential systems; E2E uses road conditions and driver steering, acceleration and braking data in an AI model.

SUM plans to conduct a remote-driving demonstration with the PV5 at Pangyo Techno Valley in Gyeonggi Province this year and validate AI training results in a real vehicle. CEO Hyun Young-jin said that building real-road experience with different vehicles and improving technology with AI determine autonomous-driving competitiveness. The company said it aims to advance its technology, including obtaining a temporary operation permit in 2027 for a vehicle using E2E autonomous driving, and grow into a physical AI mobility company.