
CEO interview: the Irish 4D imaging radar taking on LiDaR

Barry Lunn has come back from Las Vegas with a smile on his face.
The CEO and co-founder of Provizio in Shannon, Ireland, has been showing its low cost 4D imaging radar system to car makers and automotive equipment suppliers at the Consumer Electronics Show (CES).
“We are unique in having a vehicle driving around with a demo and we have two JLR vehicles on the road in Ireland with a 1km cocoon and the technique we developed allows us to get more range and resolution out of every radar pulse. It’s a really low cost way of getting high resolution radar data which we demonstrated in Las Vegas,” he said.
That Prime radar system has a 600m range for the front radar and 300m on each side and the rear, creating a dense point cloud of data for a square kilometre around a vehicle.
- Provizio 6K perception radar creates 1km2 safety cocoon
- Automotive trends at CES 2024
- ‘Single chip’ radar sensor for automotive satellite architectures
“We consider ourselves a tier 1.5 supplier as we allow the car company to get involved in the downstream design to solve their problems,” he tells eeNews Europe. “Provizio Prime allows small object detection at scale and works in all weathers so its super complementary to cameras and gives the OEM L3+ without LiDaR. In million unit quantites the price drops below $100.”
“We have a lot of insight into Lidar and one of the things that never gets talked about in reducing the form factor and price point to $500. You can put a radar into a bumper for the cost of a screw. For automotive people the lidar costs $2000 to integrate for cooling and the noise as well as the $800 cost of the sensor.”
To achieve this lower cost the Provizio system uses a high volume radar chip from Texas Instruments with a patented active antenna. “That’s what allows us to get the extended range and resolution,” he tells eeNews Europe. “Our front end is planar antenna with our own ICs so its super cheap to make.”

Barry Lunn showing the point cloud of the Provizio software defined 4D imaging radar on test on the roads around Shannon, Ireland
The company is on its second generation of front end chips, including a low noise amplifier in the 76 to 81GHz made at Global Foundries. These are surface mount, fully packaged parts that can be easily integrated into the bumper of a vehicle with the antenna.
“We are doing interesting things in the antenna layout but not doing complex things in the topology to keep the cost down. On the receive path the chip is a LNA and that reduces the noise floor, rather than increasing power which adds noise.”
- Saab opens radar plant as UK Centre of Excellence
- Reference design for first 28nm short range automotive radar
- World’s first 240GHz radar transceiver in silicon
This is key to the modular approach for the 4D imaging radar that separates out the antenna, radar, processing, software and classification of the data.
“We made everything modular so an established Tier 1 can use the active antenna and ICs with their own radar to get better resolution. Then we build the radar for customers.”
“Then it’s the perception stack and we have our own AI team and are building up our own neural network classification but other customers can use our point cloud for their AI if they want.”
“The really important thing is that the fact that we have perception team gives us a huge advantage in the design of the radar for perception rather than detection As the AI/ML team get deeper into the stack they want raw data. Some of the things that cause problems such as side lobes they are experimenting with reducing these at the signal level and the output.
“Our point cloud is 40 to 50x denser than a standard antenna, that’s huge in perception and classification,” he said. “We use generative AI to improve the point cloud by 4 to 10% but that’s not improving the resolution, that’s improving the inferencing post DSP .
“We are building a perception stack and safety stack. The camera stack will always be there. The fusion stack is in between and that’s why we have a fusion team and the three streams come together to deliver autonomous driving.”
- Materials boost for ADAS radar
- Radar crucial for autonomous driving, studies say
- NXP backs Zendar for 4D radar
The processing for the software-defined 4D imaging radar can be handled on a range of chips, which is important for the code base.
“We are embedded in the Nvidia ecosystem so that leads to their partners and with Drive Orin. That’s how Nvidia sees the world but OEMs can see it differently. We can run the processing wherever the OEM wants but with a single 4 x 4 TI chip we can deliver the same performance as the current 6 chip design and that has a huge impact on the cost.”
“We also have the NXP roadmap and our decision was early to go with TI but we are agnostic about the processing. It comes down to the choices of the OEM. A number of the projects have an established Tier One at the table and the product will be mass produced by them, so then managing the code base becomes their problem. That’s why the software layer needs to be separated from the hardware,” he said.
Provizio has projects with over 10 OEMs from top tier car makers to agricultural and mining, he says. “We are going through the productionisation process with Mergon building systems to order to generate revenue for Q2 this year for industrial applications. The next stage is the big Tier One partnerships and we are working on a bunch of those right now.”
Back from CES, and the US is a big opportunity for the Irish automotive technology company and its 4D imaging radar.
“The big opportunity is America where we are getting a lot of traction. The other one is Japan as they have skipped a generation in 4D imaging radar so that is proving really interesting.”
