Argo AI Lidar Sensor: Enabling Scalable Self-Driving Tech

Argo AI's Lidar Acquisition and its Impact on Autonomous Vehicle Development

Four years prior, Argo AI completed its initial acquisition as a burgeoning self-driving vehicle company with substantial backing. The company now asserts that the purchase of lidar specialist Princeton Lightwave is yielding positive results. This acquisition is expected to facilitate the deployment of autonomous vehicles capable of commercial operation on both highways and within densely populated urban settings, beginning in the coming year.

New Long-Range Lidar Sensor Details

Argo AI revealed specifics on Tuesday regarding a long-range lidar sensor. This sensor is designed to perceive objects up to 400 meters away, delivering high-resolution, photorealistic imagery. It also boasts the capability to identify dark and distant objects exhibiting low reflectivity.

Currently, the initial production run of these lidar sensors is integrated into a portion of Argo’s test vehicle fleet. This fleet presently consists of Ford Fusion Hybrid sedans and Ford Escape Hybrid SUVs. By year's end, Argo intends to transition its entire test fleet – approximately 150 vehicles – to the Ford Escape Hybrid model, each equipped with this internally developed lidar sensor.

Commercial Deployment Plans

Ford, a key investor and customer of Argo AI, is planning to implement autonomous vehicles for ride-hailing and delivery services in 2022. Volkswagen, another investor and customer, anticipates launching commercial operations by 2025.

Cost-Effectiveness and Scalability

According to Argo CEO and co-founder Bryan Salesky, the technical specifications of the lidar sensor are not the sole consideration. The sensor was engineered to be both cost-effective and suitable for large-scale manufacturing – crucial factors for any company aiming to commercialize autonomous vehicle technology.

Salesky highlighted a gap in the market when Argo was founded. While Waymo had developed long-range lidar sensors, they were not commercially available for other developers. “We strategically chose to acquire a company to address this long-range requirement,” he stated. “This decision has proven transformative for our self-driving system, accelerating our progress to the point where we are now integrating the sensor into vehicles and expanding testing into both urban and highway environments.”

The Importance of Lidar Technology

Lidar, or light detection and ranging, utilizes laser light to measure distances and create a precise 3D map of the surrounding environment. It is widely regarded within the industry as a vital sensor for the safe and widespread deployment of autonomous vehicles.

Currently, over 70 companies are engaged in lidar development, with some recently becoming publicly traded through mergers with special purpose acquisition companies. Each company asserts technical advancements and cost advantages. Furthermore, autonomous vehicle developers like Cruise and Aurora, similar to Argo, have acquired lidar companies to foster in-house solutions, gaining a competitive edge and reducing reliance on external suppliers such as Velodyne.

Ford's Shift in Strategy

Ford, having invested $1 billion in Argo AI, had previously supported lidar manufacturer Velodyne, a leading supplier in the market. However, advancements within Argo have prompted a change in Ford’s approach.

Veoneer announced in 2019 a collaboration with Velodyne to supply sensors to an unidentified autonomous vehicle customer. In February, Veoneer reported losing this contract. Speculation suggested Ford or Argo AI were the customer. Subsequently, Ford disclosed in a regulatory filing that it had divested its 7.6% stake in Velodyne, reinforcing its commitment to Argo’s internally developed lidar technology.

Potential Performance Advantages

Sam Abuelsamid, principal analyst at Guidehouse Insights, commented that, “If the sensor performs as claimed, it should offer a significant performance improvement over Velodyne’s offerings and provide greater operational safety at highway speeds.” Abuelsamid identified factors like wavelength and sensitivity as potential areas where Argo’s sensor could provide a competitive advantage.

• Key Technology: Lidar is crucial for creating detailed 3D maps.
• Commercialization: Cost-effective manufacturing is vital for widespread adoption.
• Strategic Acquisitions: Companies are investing in in-house lidar development.

From City Streets to High-Speed Highways

Argo AI in the Strip District on Monday April 26, 2021 in Pittsburgh, Pennsylvania. (Photo by Jared Wickerham/for Argo AI)

Lidar sensors emit millions of light beams or pulses per second. These beams are used to identify surrounding objects and measure the time it takes for the light to return. This data creates a detailed point cloud, effectively a 3D image of the environment.

The point cloud generated by the sensor reveals the objects present and accurately calculates their distance. Argo’s sensor technology utilizes what is known as Geiger-mode time of flight lidar. This system employs beam detectors capable of registering even the faintest particle of light.

Argo asserts that its single-photon sensors can construct images of objects with low reflectivity, such as a vehicle with a black paint job, from a significantly greater distance. This is in comparison to traditional linear time of flight lidar systems.

Furthermore, Argo’s lidar operates at a wavelength exceeding 1400 nanometers. This wavelength theoretically allows for increased power output, contributing to extended range capabilities.

Industry analyst Sam Abuelsamid points out that conventional 905 nm lidars typically have limitations. They are generally effective for vehicles traveling at speeds of approximately 40 to 45 mph. This suggests Argo’s sensor could function reliably at highway speeds.

“The utilization of Geiger-mode photodiodes and pixel binning within Argo lidar significantly enhances its sensitivity,” Abuelsamid explained. The ability to detect single photons and then employ software for statistical analysis to filter noise appears to be a key advantage.

Detecting low-reflectivity objects, like tire treads on trucks or very dark colored vehicles, is crucial for accurate perception. Argo’s system is designed to address this challenge.

The Argo sensor features a mechanical spinning design, a common approach originating with the Velodyne HDL-64. However, Argo’s implementation rotates the sensor’s outer surface.

This design choice is intended to effectively shed water from the sensor, maintaining its cleanliness and optimal performance. A clean sensor ensures consistent and reliable data acquisition.

The combination of these technologies will enable Argo to create a self-driving system adaptable to diverse scenarios. This includes navigating slow-speed, densely populated urban areas.

It also encompasses higher-speed boulevards with mixed traffic – pedestrians, cyclists, and cars – and finally, high-speed highway driving. The system is designed for versatility.

Argo has focused its testing efforts on urban environments. Key cities include Austin, Detroit, Miami, Palo Alto, Pittsburgh, and Washington, D.C.

However, Volkswagen, a recent investor and customer, is also keenly interested in autonomous driving capabilities on highways. Argo intends to expand its testing program to additional cities this year.

Munich is among the cities slated for inclusion in the expanded testing phase. This expansion reflects Argo’s commitment to developing a comprehensive self-driving solution.

Manufacturing Collaboration and Potential Uses

For over a year, Argo has collaborated with a contract manufacturing firm specializing in optoelectronics assembly. Production is projected to reach hundreds of sensors by year-end, with subsequent scaling planned. The identity of this manufacturing partner remains undisclosed by Argo.

The self-driving system developed by Argo is designed for versatility, enabling its deployment across diverse business frameworks. While Ford and VW, Argo’s primary clients, will ultimately determine the specific applications, current focus lies on robotaxis and middle-mile logistics.

However, Salesky indicated that the technology also holds potential within the trucking sector. “Our initial concentration is on the transportation of goods and people,” he explained, “but trucking is receiving significant consideration.”

Although not yet a priority, Argo is actively evaluating the feasibility of applying its technology to trucking applications. Salesky believes this area presents a compelling opportunity for their innovations.

Beyond internal use, Argo envisions licensing its long-range lidar technology for broader applications. The core technology can be adapted into various sensor configurations, creating new product possibilities. “This represents a potentially valuable licensing opportunity,” Salesky stated, emphasizing that autonomous vehicle development remains the primary focus.

While immediate sales within the automotive industry are not planned, Salesky acknowledged the potential. He further suggested that the underlying technology could be adapted to serve industries like mining, agriculture, and oil & gas.