NVIDIA's (NVDA -2.28%) artificial intelligence (AI) products for data centers and autonomous cars get significant press coverage, but its drone AI products and projects don't get the attention they deserve.
NVIDIA's graphics processing unit (GPU)-based approach to AI is being rapidly adopted to smarten up drones for a wide variety of applications. This is good news for investors, since the global civil drone market grew about 36% last year, according to research firm Gartner, and is widely expected to balloon in size in the coming decade.
In October, I wrote about NVIDIA's AI being used by Chinese e-commerce giant JD.com to power its drone delivery initiative, and by Avitas Systems, a General Electric venture, for its drone industrial inspections. Last month, NVIDIA scored yet another win in the drone space for an entirely different application: monitoring highways for compliance with high-occupancy vehicle/high-occupancy toll (HOV/HOT) lane regulations.
Here's what you should know.
Drones as carpooling-lane traffic cops
In November, the team of NVIDIA, IBM, the University of Toronto, and The Sky Guys, which is Canada's leader in drone solutions and technology, was awarded $750,000 from Ontario Centres of Excellence, as part of the Small Business Innovation Challenge (SBIC). The "group will work closely with the Ministry of Transportation of Ontario (MTO) to develop an artificial intelligence-enabled drone solution to monitor Ontario's 400-series highways in response to the Vehicle Occupancy Detection Problem Statement," according to The Sky Guys' press release.
This crux of the referenced "Problem Statement" is as follows: "As high-occupancy toll (HOT) lanes are implemented in Ontario and abroad, there is a growing need for the development of an enforceable means to determine if a vehicle in a specific lane should be billed and/or fined based on the number of individuals in the vehicle." HOT lanes are ones in which carpooling drivers would continue to drive for free, but other drivers would pay a toll.
As background, the 2015 Ontario budget includes that the province would continue to assess the feasibility of building new HOT lanes and converting select HOV lanes into HOT lanes. The HOT Lane Pilot is the first step in Ontario's plan to implement HOT lanes throughout the province in 2020/2021. The pilot began in September 2016, with the designation of a 16-kilometer stretch (or nearly 10 miles) on the Queen Elizabeth Way (QEW) -- a highway that links Toronto with the Niagara Peninsula and Buffalo, New York -- as a HOT lane. Ontario will use this section of the QEW to test new tolling technologies over the two to four years the pilot is expected to operate. Findings will help the province in its long-term planning for future HOT lanes.
Big growth potential if hurdles can be addressed
A $750,000 project divvied up in some unknown way between the four partners means that NVIDIA won't be taking home a king's ransom for its work along a queen's highway. However, there's big growth potential in using drones to monitor carpooling lanes. In the United States alone, there are more than 300 kilometers (about 186 miles) of toll lanes that provide occupancy-based discounts, and more than 5,000 additional kilometers (3,107 miles) of HOV lanes, according to The Sky Guys.
A vehicle occupancy detection technology that would automatically provide billing and regulation enforcement support could potentially be implemented along these stretches of roads in the U.S., as well as in other countries that have HOV/HOT lanes. Such a technology has the potential to increase compliance levels and toll revenue, while decreasing the cost of law enforcement labor costs. Moreover, drones have other possible traffic-cop-like applications beyond those involving HOV/HOT lanes, as well as the potential to assist law enforcement in areas other than on roads and highways.
In order for this drone pilot program to get off the ground, privacy concerns will need to be adequately addressed, as the drones will reportedly take several images of a vehicle. The Sky Guys has said its software will obscure identifying details, according to The Globe and Mail, though whether this will be enough to satisfy critics remains to be seen. Moreover, currently drones are not permitted to fly out of visual range of their operator, per Transport Canada regulations. Even if these two hurdles aren't adequately addressed and the drone traffic cops don't take flight, the mere fact that NVIDIA was chosen by Canada's leading drone solutions and technology company to be a part of this project speaks volumes about the company's growth potential in the drone space.