Inside GNSS Logo

Kodiak Robotics is based on lightweight Autonomous Truck Mapping PNT – Inside GNSS

Kodiak Robotics, Inc. Recently engaged in the development of automated solutions for long-haul truck roads in the southern parts of the United States

The self-driving trucking company made a big announcement on that front in August, revealing a partnership with Pilot, North America’s largest travel hub operator, to develop autonomous trucking services at travel hubs Pilot and Flying J.

Companies are setting up an independent truck port in the Atlanta area to evaluate potential service offerings and explore scalable solutions. Possibilities include spaces to pick up and drop off standalone truck loads; carrying out inspections; Truck maintenance and refueling. The ability to transfer data for feature development and mapping.

The partnership is important to both Kodiak and the industry. Kodiak CEO Don Burnett said the company is placing players like Pilot and its travel hubs as key locations to facilitate the various services that autonomous trucks will need when they are in production and are commercially deployed. In addition, the pilot centers will serve as access points for data transmission.

The Pilot partnership is just the latest development in Kodiak’s exponential growth phase in 2022, with a major expansion coming in its range of services and partner network as well.

In July, the company announced a partnership with 10 Roads Express, a company that provides time-sensitive ground transportation for the United States Postal Service, to expand the company’s service to Florida. And earlier this year, Kodiak announced a new route between Dallas and Oklahoma City with CEVA Logistics and a route between Dallas and Atlanta with US Xpress.

Fourth generation independent truck

This commercial success is driven in part by technology pioneering in the fourth generation Kodiak autonomous truck. The new generation is designed to improve the durability of the autonomous system, with greater fleet uptime, manufacturing and serviceability — all essential to scaling technology quickly, safely and efficiently, according to the company.

“Complex, bulky systems that require an engineer to manually build and manually tune are expensive, unreliable, and difficult to debug,” said Burnett, who co-founded Mountain View, California-based Kodiak Robotics with COO Paz Eshel. “We believe reliability and scalability stem from simplicity, and the best hardware modifications should be barely visible. Our 4th generation platform is designed for simple, scalable production, which means easy calibration, troubleshooting, and maintenance for our partners.”

The truck features a standard and more discreet sensor array in three locations – a slim ‘center pod’ on the front roof over the windshield and integrated bags in both side mirrors. The better integrated sensor placement is said to greatly simplify sensor installation and maintenance while also increasing safety.

The autonomous driving system features Luminar’s Iris LiDAR, Hesai’s 360-degree scanning LiDAR for side and rear vision detection, ZF full-range radar, and Nvidia’s Drive platform for AI brains.

The Kodiak Vision Perception System considers every sensor — including LiDAR, camera and radar — essential, according to the company. The three sensors are specifically designed to meet the needs of autonomous trucks, which must “see” long distances in a variety of weather conditions to operate safely at highway speeds.

The system integrates information from sensors and takes into account the relative strengths and weaknesses of each type. It includes additional redundancy and data validation, adding another layer of security to the autonomous driving system.

The hanging mirror chamber – which will start with one long-range Hesai LiDAR radar and three cameras – does not require specialized sensor calibration. Instead of replacing the sensor that needs maintenance, a mechanic can replace the mirror case in minutes. This single point of integration will allow maintenance and serviceability at scale.

To make sense of all the data, Trucks will feature Nvidia Drive Orin, once available, as the supercomputing platform. With more than 250 TOPS (Trillion Operations Per Second) of computing performance, the safety platform is designed and addresses systematic safety standards such as ISO 26262 ASIL-D (Automotive Safety Standard-D). In the meantime, Kodiak will use the current generation Nvidia Drive AGX Pegasus to process data from the cameras.

Kodiak_Robotics_was_co-Found_by_l-to-r_CEO_Don_Burnette_and_COO_Paz_Eshel.
Kodiak Robotics was co-founded by CEO Don Burnette and COO Paz Eshel.

Importance of PNT

When it comes to positioning, navigation, and timing (PNT), Burnette said it’s “definitely an area where we feel Kodiak is really innovating in the space.”

Within the Mapping and Localization framework, he said, “Ultimately the bot needs to answer the question ‘Where am I?’ And once she knows where she is, she needs to ask, “How do I drive from here?” And then just repeat those questions.”

Historically, most companies have implemented high-definition (HD) maps of the environment using vehicle sensors to identify minute details such as road texture surfaces, paint marks, tree trunks, buildings, building sides, etc., Burnett said.

“You name it, they put it on a map, and then they use that map to put themselves in a very precise position in the real world,” he said. And then they use IMU [inertial measurement unit] to satisfy these position-based preferences.”

His company differs from the industry in this respect.

“We do it differently,” he said. “We have a very sparse mapping solution that only includes the road network – lane connection information.”

For example, the Kodiak system and a lightweight Sparse map keep track of the number of lanes and their relationship to each other, knowing where the exits are and
Alfalfa leaves.

“From there, we translate based on what our sensors see for driving-related lane markings, in much the same way. [that] “People do,” he said.

Kodiak engineers use the IMU to interpolate the location of the truck as it moves along the road.

“Our positioning is very high accuracy in the lateral sense, but it is not very high accuracy in the longitudinal sense,” Burnett said. “It just doesn’t matter if we’re within a foot or foot of the road. As long as we’re close enough to be around, we can spot key signs telling us where we need to get out and where we expect other vehicles to be.”

This is where GPS comes in.

“We have a very loose reliance on GPS just to boot the system when we just start to kind of tell us where we are initially,” he said, “but also then to gradually pull us longitudinally to maintain that near-accuracy.”

Kodiak_focuses_on_easy_SensorPod_maintenance.

Reliability is king

While performance and cost are important considerations for IMUs and GPS units — as well as perception sensors — for autonomy, the primary metric that Kodiak developers are concerned with is reliability.

“I think this is a bit of a surprise to most people, and it applies to all sensors,” Burnett said. “At this point, we are not looking for better performance. I think we have the performance we need from our sensors, computations and hardware that would be acceptable to launch this product safely. Cost is somewhat important, but what really matters to us is reliability.”

The company hasn’t announced the sourcing of its IMU and GPS, but it wants suppliers that are able to build solutions that can withstand the harsh environment trucks face day in and day out without breaking.

“If you can build a unit that will go hundreds of thousands of miles on the highway without breaking, that’s what we care about,” he said. “We care about reliability much more than any kind of fancy gadget.”

The ability to withstand typical shocks/vibrations is one of the most important considerations, especially for IMUs, along with water ingress and temperature fluctuations.

“We drive in the summer heat in Texas where it can be very hot, and he also needs to be able to work in the bitter cold,” he said. “So temperature, shock, atmosphere, water, ingress, reliability — just general wear and tear — these are the kinds of things we evaluate.”

These assessments continue as the company aims to integrate its self-driving software and hardware, Kodiak Driver, into production customer trucks in early 2025. Kodiak Driver will operate self-driving fleets for a low subscription fee per mile.


#Kodiak #Robotics #based #lightweight #Autonomous #Truck #Mapping #PNT #GNSS

Leave a Comment

Your email address will not be published. Required fields are marked *