Mr. Laroche, you are CEO of Jakarto, a Quebec turnkey mapping company that has developed an innovative and cutting-edge software for exploring and measuring the urban landscape.
Tell us a little about your company and the services it offers.
Jakarto is a high-definition mapping company. Our process allows the three-dimensional (3D) digitization of a given place, for example a municipality or a road network. Paired with GPS localization, this allows an accurate representation, within a few centimeters, of a physical territory. We produce what is called a "Digital Twin."
Data capture is done using vehicles equipped with five ultra-high-definition cameras and two high-precision Lidar scanners, in order to determine as precisely as possible the distance between objects. The data thus collected are processed by artificial intelligence algorithms enabling the recognition of physical objects.
Jakarto plans at minimum one pass per year on each street of a municipality. In the case of construction work or another indicator of change, additional passes are added.
You previously held a senior management position in a thriving company. Why did you decide to launch into the field of mapping and intelligent mobility?
The idea to found Jakarto sprang up two years ago. The development of this new project began with the resources of JLR, the company I had worked for for 20 years. The potential of the project, but above all the scale of the task, quickly became apparent. I therefore decided to devote myself to it full-time and to give up the position of CEO to my partner.
Starting a company from scratch is an incredible personal challenge. You have to build a new ecosystem and recreate your network of contacts, but it's very stimulating! Two years later, we have already hired eight employees.
How does a young, modestly sized company like Jakarto distinguish itself from the giants present in the mapping ecosystem, such as Google or Apple?
The degree of precision of our data is the best on the planet at the moment. Google and Apple are not yet doing three-dimensional mapping. Apple has begun developing capture units similar to ours, but the equipment they use is less performant and, therefore, less precise.
Moreover, these big players are mainly targeting a mass market. On our side, we aim for a more niche clientele. Jakarto will, for example, deliver 3D maps to about twenty cities in the coming weeks. Our real competitors are companies like the Dutch group TomTom or the mapping company Here. But we are certainly making our place in the market.
Is 3D mapping important for the development of intelligent mobility?
Put together an ultra-precise 3D map integrated into navigation systems, complement it with very high-precision cameras and artificial intelligence algorithms to recognize signage, and there you have your autonomous vehicle!
3D representations of the road network, like the signage recognition algorithms we are developing, are essential innovations for the rise of autonomous driving. I think that in the near future, mapping will be integrated into communication systems between vehicles and street furniture, for example a lamppost or a stop sign. To allow the update of a territory's map in real time, data will ideally be automatically uploaded into navigation systems when vehicles pass near stations provided for this purpose.
What challenges still need to be met in 3D mapping with regard to intelligent mobility and autonomous driving?
We need to perfect our signage recognition algorithms. Six or seven classes of physical objects, such as pedestrians, cars and street furniture, can already be identified by our robots. However, we want to arrive at the identification of many more elements, such as lane markings, reserved lanes, etc.
But our biggest short-term challenge remains the ability to scale up. It's easy to manage data collected over 10 km of road, but the task is much more complex when you want to map hundreds of thousands of kilometers! We then encounter problems of hosting, distribution and compression of information.
What are your priorities for the coming years?
Many projects are currently underway. In fact, a capture unit will soon set out to map the entire Trans-Canada Highway and will make the collected data public. I think that's going to make some noise!
We have a few projects with Université Laval, including participation in Canada's Smart Cities Challenge.
Offices will be opened in the coming weeks in Waterloo. There we will develop maps in partnership with the university in that city.
Discussions are underway with an automaker to provide mapping for Canada, and perhaps even for all of North America. It's very exciting!
The video game market also seems very promising and one we would eventually like to penetrate. The software we use for map visualization is, moreover, the same as that used by many video game developers.
In short, there is no shortage of prospects! We don't sit idle at Jakarto!
