Baidu Announces 'Apollo Lite' a Camera-based Autonomous Driving Technology

author: Eric Walz   

A Baidu Apollo self-driving car parked outside the company's headquarters in Beijing.

Baidu announced its camera-based Apollo Lite this week. Apollo Lite is a vision-based autonomous driving solution that leverages multiple cameras to achieve Level 4 (L4) autonomous driving. Similar to how Tesla Autopilot works, the camera-based computer vision system does not rely on lidar.

Apollo is Baudi's open autonomous driving platform. The internet search giant is working with over 120 global industry partners on the Apollo platform. The collaboration with partners is designed to speed up the development of autonomous driving technology.

Baidu says that Apollo Lite is China's only vision-based Level-4 autonomous driving solution in use today. It has the capacity to process vast amounts of data generated from a suite of 10 cameras. The cameras can detect objects as far as 700 feet away. The system provides real-time, 360-degree sensing of the environment around the vehicle.

"A robust vision-based system is critical to the safety of autonomous driving, especially in high-speed situations where real-time sensing is critical," said Liang Wang, head of Apollo's technical committee. "Apollo Lite further strengthens Baidu's sensor fusion based L4 autonomous driving system that leverages the capabilities of camera, Lidar and radar to achieve the ‘true redundancy' necessary for a safe and fully autonomous driving experience."

Perceptions systems, such as lidar and cameras, are one of the most important technologies for self-driving cars and advancements in perception systems have sped the development of autonomous driving. Compared with the current laser radar solution in the industry, the visual perception solution using cameras is also much less expensive and easier to design.

Tesla CEO Elon Musk is a proponent of camera-based systems due to their lower costs. Advanced lidar systems from companies such as Silicon Valley-based Velodyne can cost tens-of thousands of dollars, much too expensive for a mass-production vehicle. Musk said that these systems are "expensive sensors that are unnecessary."

Vehicle outfitted with Apollo Lite have already successfully completed L4 autonomous driving on public roads in Beijing, without having to rely on lidar. The vehicles are capable of operating autonomously within specific areas and under certain weather conditions.

Since launching the open-sourced the Apollo software platform in 2017, Baidu is helping developers and partners to quickly build their own autonomous driving systems.

Five months ago, Baidu introduced its Apollo Enterprise, a commercial initiative that supports a suite of customizable autonomous driving products, such as automated highway driving and autonomous parking for mass-production vehicles.

Baidu said its latest version, Apollo 3.5, uses over 390,000 lines of code from 12,000 Github contributors.  

Eric Walz

Originally hailing from New Jersey, Eric is a automotive & technology reporter covering the high-tech industry here in Silicon Valley. He has over 15 years of automotive experience and a bachelors degree in computer science. These skills, combined with technical writing and news reporting, allows him to fully understand and identify new and innovative technologies in the auto industry and beyond. He has worked at Uber on self-driving cars and as a technical writer, helping people to understand and work with technology.