Baidu’s OpenEdge Platform Could Speed Development of Autonomous Applications

Chinese tech giant Baidu is launching its OpenEdge open source edge computing platform at this week’s CES expo in Las Vegas, which could have important applications for autonomous and connected vehicles.

Moving the computing process closer to the source of the data can greatly reduce latency, lower bandwidth usage and ultimately bring real-time and immersive applications to users.

“Considering the large number of vehicles that can operate on road networks, having the ability to manage these connections in real time is essential,” Sam Barker, senior analyst for Juniper Research, told TU-Automotive in an email.

By providing an open source platform, Baidu has also made it easier for developers to create their own edge computing applications.

Meanwhile, open sourcing enables a greater number of developers to work on multiple projects simultaneously, accelerating the time to market for many.

“In a market that is still very much focused on development, this strategy is driven by opening up development to the highest number of developers as possible, including development and testing of products,” Barker said.

OpenEdge is the local package component of Baidu Intelligent Edge (BIE), which consists of a cloud-based management suite.

The management suite provides functions to manage the edge nodes, the edge apps, and the resources such as certification, password, and program code, allowing BIE installed apps to take advantage of both cloud reliability and edge locality.

Devices deployed with BIE can also exchange data with ABC Intelligent Cloud, perform filtering calculation on sensitive data, caching data and providing real-time feedback control when the network can’t be found or is unstable.

Barker explained the key word for maximizing the safety of autonomous vehicles is latency. Although 5G networks are helping to reduce the latency of automotive cellular connections, edge computing network infrastructures will further reduce latency of automotive connections.

Additionally, edge computing will enable the processing of data closer to the edge of the network, which reduces strain on cellular networks carrying any automotive data.

“As vehicles become increasingly connected, the data generated per vehicle on a daily basis is going to experience significant growth,” Barker said. “Edge computing architecture will enable this processing of data from the edge of network with only the insight gained sent to the network core.”

Baidu also introduced two products based on its intelligent edge computing technologies, BIE-AI-Box and BIE-AI-Board.

A joint launch with chipmaker Intel, BIE-AI-Box is a hardware unit incorporating Baidu technologies designed for in-vehicle video analysis.

The unit connects with in-vehicle cameras to optimize video analysis and will also provide AI apps for road recognition, car body monitoring, and driver’s behavior recognition.

The second product is BIE-AI-Board, launched in partnership with Dutch semiconductor manufacturer NXP. The software installed onboard can be embedded into cameras, drones and robots to perform a variety of detection tasks.

One application would be the measurement of charging stations for new energy vehicles, where the equipment can collect and analyze data from each charge.

This would help proprietors assess the status of the charging equipment and report relevant data to station and vehicle owners, potentially lengthening the life cycle of charging equipment.

— Nathan Eddy is a filmmaker and freelance journalist based in Berlin. Follow him on Twitter.