Kurtis McBride describes what he does for a living in simple terms: “Make the traffic suck less a little bit more every day.”
Everything in his world comes down to traffic lights. The 39-year-old founder of Miovision readily catalogues the litany of frustrations associated with these fixtures of city life. There’s the “split,” which is what happens when you’re stuck in the left-turn lane and the light changes. And “co-ordination failure,” when you hit every red light. And then there’s “conflict,” a catch-all term for when the signal is somehow implicated in a collision (or near-miss) between a car and a cyclist or pedestrian.
Traffic engineers traditionally tried to adjust traffic light intervals by manually counting vehicles and then doing the math (tweaking intersections is the ultimate zero-sum game—more time for one direction means less for the other). McBride automated that cumbersome task, feeding data back to municipal traffic control centres and transportation planners.
He founded his Waterloo, Ont., tech firm in 2005 before even graduating from university. During the first seven years, Miovision’s technology became the “de facto standard way to do vehicle counts in intersections across North America,” McBride says. The product can now be found in over 60 countries.
Then, three years ago, Miovision raised $30 million from a syndicate of Waterloo-based investors and three venture capital firms. The goal: to develop sophisticated computers that live next to intersections (inside those upright boxes you’ve likely seen) and are linked to digital cameras programmed to see and interpret what is happening in the intersection. TrafficLink, which can adjust signal intervals automatically depending on traffic conditions, is equipped with artificial intelligence, computer vision software and the ability to communicate with signal controllers installed at adjacent intersections.
“You’re putting a lot of technology into the intersection,” McBride says, adding that Miovision has installed 2,000 TrafficLinks to date. He claims the firm’s technology can accelerate the transportation network by 20% to 40%. The market seems impressed. The company, which employs about 200 people in a funky converted tire warehouse in Kitchener, Ont., raised an additional $15 million last March.
Unsurprisingly, Miovision markets itself as a “smart city” company. Its promotional materials and branding have a decidedly urban feel, and its mission reflects the smart city hype: Marrying technology, data and municipal systems can improve the quality of life in cities.
While estimates vary widely, the smart city sector is vast. The IMARC Group, a New York market research firm, recently estimated that the global market was worth over US$312 billion in 2018 and appears to be on track to produce a compounded annual growth rate of 17.6% through 2024. “This has become such a hot business area,” says Matthew Siemiatycki, a geographer with the University of Toronto and Canada Research Chair in infrastructure finance. “The money is flowing into this space like never before.”
The players include everyone from tech startups around Tel Aviv to multinationals like Cisco and Siemens. Alphabet’s smart city subsidiary, Sidewalk Labs, has whipped up controversy since alighting on Toronto’s post-industrial waterfront in October 2017.
Siemiatycki also points to firms in adjacent markets that hold great potential for municipal governments. GeoTab, in Oakville, Ont., has developed an asset-management tool for fleet operators. The firm’s vehicle-based GPS sensors, which are connected to its platform, originally allowed delivery firms to optimize their routes, potentially boosting profit margins in the highly competitive freight sector. But the devices have been upgraded so they can also listen to engine performance, using an AI application to anticipate maintenance needs before vehicles actually break down. All municipal fleets—such as buses and police cruisers—could benefit from this kind of care, Siemiatycki says.
Indeed, the workhorses of the smart city industry are sensors that can be deployed anywhere to gather and transmit all manner of data. They come in lots of flavours. Some are as simple as compact WiFi-connected devices that do nothing more than determine when a waste bin needs to be emptied, while others are as complex as Miovision’s high-end intersection cameras, which are programmed to “see” the difference between a pedestrian and a cyclist, and can account for visual occlusions, like signs, newspaper boxes and other non-vehicular detritus. Some sensors do just one thing—measure water flow, for example—while others, such as GE Current’s CityIQ nodes, are bundled to perform all sorts of tasks, including automatically adjusting LED street lights to ambient light conditions, measuring air pollution and collecting video.
Millions of sensors have already been installed in cities around the world. But as sensors, even sophisticated ones, become commodity items, the deployment will accelerate. What could turbocharge this already robust industry is the much-hyped but little understood expansion of 5G wireless networks, something that will likely take place this decade. Peter Linder, head of 5G marketing in North America at Ericsson, says the U.S., Japan, China and South Korea are moving fastest, while Canada’s commercial 5G network isn’t expected to begin construction for another 18 to 24 months.
The premise behind 5G is that it promises to significantly increase data speeds and reduce lag times, allowing systems to communicate with one another wirelessly at a speed that is currently only possible with broadband fibre optic networks. What’s more, 5G technology is intended to rely on a mostly unused portion of the radio frequency spectrum. The reason for this approach has everything to do with overcrowding elsewhere on the dial. But the logistical wrinkle—and it’s huge—is that 5G networks require a much denser arrangement of cell towers or transmitters than currently exists.
In fast-moving jurisdictions, like the U.S., regulators have already allocated the spectrum. Meanwhile, state governments are moving quickly to prevent municipalities from banning the deployment of additional 5G transmitters, which are the size of pizza boxes. These new rules come in response to opposition from neighbourhood groups and people who fear the proliferation of cell towers. “There’s a lot of angst around it,” observes Anthony Townsend, an American smart city consultant and author. “Even people who don’t think they cause cancer are nervous.”
Here, the Canadian Radio-Television and Telecommunications Commission hasn’t yet auctioned off 5G frequencies. While various research projects are underway, involving various governments and telecom giants like Ericsson and Huawei, a commercial launch remains a ways off. As Tejas Rao, Accenture’s managing director for 5G in the Canadian market, says, “I kind of look at 2020 as the starting point.”
The reason for the slower pace doesn’t just involve regulatory and technical hurdles; it’s also about identifying commercial applications that require this level of service. In big cities, 4G and LTE networks are more than capable of providing streaming services. Linder points instead to smart city “use cases” that would work most effectively on a 5G network. He cites examples of how it could be used for security, such as CCTV cameras trained on public spaces like arenas. Existing CCTVs feed low-resolution video back to a control hub, where they may be monitored by human beings. If connected to a 5G network, they could not only deliver high-resolution streaming video but also layer on analytics, such as algorithms that anticipate disturbances, and notify authorities. He adds that drones designed to operate via a 5G video feed could be sent to fly over inaccessible accident sites or fires to provide real-time information about what’s going on.
Townsend adds that vehicle-based wireless devices, including digital cameras, could eventually become the “killer apps” for 5G, with on-board systems that allow cars, trucks and buses to communicate with one another as they navigate city streets. Of course, such applications gain even more relevance with the advent of autonomous vehicles, which will depend on on-board sensors, as well as GPS and other wireless connections. Some fleets could eventually be operated remotely, with one “driver” in charge of several vehicles. Transportation applications, Townsend says, “could quickly become the bulk of the traffic on the network.”
Yet all these far-reaching ideas continue to be years away from reality. It remains to be seen how the telecom carriers—who have to make the upfront capital costs in building out the 5G network—will recoup their investments. “Today, a lot of this is in test beds,” says Rao, citing a testing facility in Ottawa and a five-city, five-year joint 5G venture between Ontario, Quebec and several tech firms.
Even without the power of 5G, the rapid deployment of sensors in the public realm has raised concerns about privacy and intrusion. Siemiatycki parses the smart city sensor world into two broad categories: those that are affixed to municipal infrastructure and monitor how it works (water flow through pipes, energy usage in a local grid, engine performance on a city bus) and those that can detect and possibly identify individuals moving through public space (for example, China’s widespread deployment of facial-recognition cameras). Those applications involved in asset management—such as sensors that can detect leaks in water mains—likely have the best potential return on investment and generate the least amount of static.
More public-facing sensors, however, have raised concerns, as has been the case in Toronto over Sidewalk Labs. San Diego recently procured almost 4,000 smart lighting boxes to attach to street lamps as part of an LED conversion program meant to reduce energy consumption. But the devices also have on-board video cameras, air-quality sensors and audio recorders. The city’s stated intention was not only to adjust street lights but also monitor parking violations and bike lane uses. As it turned out, San Diego’s police department began using the video—five days’ worth is stored on the devices at any given time—for investigations, a development that triggered an intense backlash that consumed San Diego city council. (These devices, though connected to a network, don’t rely on 5G wireless.)
Townsend points out that a growing number of smart city sensor developers, sensitive to such conflicts, now build in software that dispenses with non-essential data to prevent the kind of mission creep that caused so much fighting in San Diego. A device meant to record pedestrian activity at a busy intersection with video might use an algorithm that blurs faces or simply registers people as blobs instead of identifiable individuals. “These companies are worried about privacy and don’t want to be seen as the bad guys,” Townsend says.
An even more likely future for smart city sensors can be found in Hull, in the U.K. Two years ago, the municipality hired Connexin, a London-based AI firm, to create what the company calls a “city operating system.” The goal was to avoid what happens in many places, where different departments end up investing in costly smart city devices or systems that can’t talk to one another, says CEO Furqan Alamgir. Hull, he says, is “agnostic” when procuring sensors; what it does insist on is that they’re all compatible with the municipality’s software platform. This, he says, allows planners to figure out new ways to combine sensors to deliver some public-policy goal.
Alamgir cites a sensor-based parking system developed in another U.K. city, Newcastle, that’s been recently rolled out. Using either CCTV or in-ground weight-detection coils, the system tracks vacant parking spaces and posts locations on navigation apps. At the same time, air-quality sensors and devices designed to track vehicle flow are continuously monitoring traffic congestion. The data is fed into a system that not only identifies conveniently located empty parking spaces but also adjusts parking pricing to generate more revenue for the city. Alamgir notes the data gathered by the sensors belongs to his client: “We are enablers, not data owners.”
Miovision, which collects and processes a huge amount of information at intersections, takes a similar approach. The data from the firm’s cameras and the analytics from its software, McBride says, ultimately belong to its client cities. “We have to sort of strike a balance between being good corporate stewards—not over-generating information but also not falling into a pattern of hubris where we tell the city what they can and can’t do with their data.”
That particular job belongs to others.
How a package can tell you that it’s damaged
by Liza Agrba
Inventory management is crucial for retailers, but there’s only so much that can be accomplished using the spreadsheets still employed by most companies. Cloud-based tracking systems can offer an improvement but, in terms of data analytics, there’s no comparing either process to having a supply chain hooked directly to the Internet of Things, or IoT.
Imagine attaching a tracking sensor to each item in your inventory that can communicate its location and temperature, and whether damage has occurred, all in real time. Data that granular would give retailers unprecedented insight into the inefficiencies in their systems. It would also cut human error out of that equation. Like many potential developments in the IoT, this one is still in its early stages, and applying it to inventory systems on a large scale, and in a cost-effective way, could still be a long way off. But its potential implications for companies’ resource efficiency are massive.
One Canadian firm working on IoT-enabled supply chains is Peytec, based in Markham, Ont. The company makes trackers that communicate an item’s location to a supply-chain management system. Peytec says its trackers are accurate within 10 centimetres and can be equipped with modular sensors that monitor variables like temperature and moisture levels. “The benefit of this system over existing ones…is that it’s much more accurate and occurs in real time,” says Peyman Moeini, Peytec’s president and CEO. “No one likes scanning items in a warehouse, which is part of why the error rate in inventory management systems tends to be so high. This technology solves a major pain point for industry.”
A porchlight that leaves itself on for you
by Liza Agrba
Your porch light turns blue to let you know it’s supposed to rain in the afternoon. Street lights autonomously adjust to weather and traffic density. A self-driving car monitors road conditions by communicating with other vehicles through its headlights and tail lights. These are a few of the potential use cases for smart lighting, an industry with a staggering range of residential, commercial and industrial applications. Indeed, the sector is projected to reach US$30.9 billion by 2025, according to Kenneth Research. Driving the growth is energy efficiency. All smart lighting systems use power-saving LED bulbs, and ambient light sensors cut energy consumption.
Nanoleaf is a Canadian example of a smart lighting firm with reduced energy use as its principle. Founded by three University of Toronto graduates, the tech startup made its mark with colour-changing lighting panels that react to music and other sounds. Now, the firm is developing smart lights that will adjust to a user’s routine without input from an app. “We developed a series of sensors that will predict your movement around a space and turn lights on and off for you,” says Gimmy Chu, Nanoleaf’s co-founder and CEO. Chu thinks the industry is moving away from app-controlled and bulb-shaped lights altogether. Instead, he says, algorithm-powered sensors and innovatively shaped fixtures-—like Nanoleaf’s own triangular panels—will come to the fore in the next few years.
The trend has colossal implications. Lights will soon do far more than help you to see. Embedded sensors are crucial to that effort, monitoring everything from external conditions to the state of the bulbs themselves. Another growing area of research is LiFi, or using light as a vehicle for data transmission. Where WiFi uses radio frequencies, LiFi transmits data using visible, ultraviolet and infrared light spectrums. Exempt from electromagnetic interference, it’s already being studied for use in hospitals and aircraft, and for vehicle-to-vehicle communication.
Every smart device needs an edge
by Liza Agrba
The Internet of Things, or IoT—connecting vehicles, cameras, light bulbs, espresso machines and more—is expected to cause a massive uptick in the amount of data transmitted across the globe, placing a heavy load on centralized data centres. And since real-time responses are part of the appeal of the IoT—especially for applications like helping self-driving cars avoid obstacles—a near-instantaneous flow of data is central to its functionality. That’s where edge computing comes in.
Edge computing involves storing and processing data as close as possible to where it’s needed. For IoT devices running on an “old-fashioned” (to push that word to its logical limit) cloud computing infrastructure, data travels to a centralized server for processing, and then a response is sent to the device. When you’re looking at connecting thousands of devices to the internet, the processing load—and resulting bandwidth costs—would be astronomical. Edge computing aims to bring storage and processing physically closer to the devices. While there will likely be a loss of computing power compared to large centralized servers, increased speed will furnish the real-time responsiveness promised by IoT ecosystems.
How to make edge computing possible is a hot research area for some of the world’s largest computing and telecom firms. In October, tech giant Intel bought a platform from Toronto-based Pivot Technology as part of its effort to double down on 5G, the next generation of faster wireless technology. Telus and Rogers have both announced major edge-computing initiatives. In partnership with California-based MobiledgeX, Telus launched an edge development program, set to open in Canadian cities in the near future. Meanwhile, Rogers hosted an edge-computing challenge at the University of British Columbia in October and recently announced a partnership with the University of Waterloo to study 5G, including edge computing.
A future without tow trucks
by Kim Hart Macneill
A truck packed with $300,000 worth of shrimp arrives at its final destination. The receiver, who fears the summer shipment entered a temperature danger zone during its journey, is hesitant to accept the load. Even five years ago, that situation could have meant a costly loss for the shipping company. But thanks to Fleet Complete’s temperature tracking system, the firm was able to prove the shellfish stayed cold for the entire trip.
The Toronto-based company’s main line of business is telematics, the merging of telecommunications and information technology, for light- and heavy-duty vehicles. Its tracking systems use wireless sensors throughout to monitor everything from location to trailer temperature control to idling time to the speed and driving styles of its employees. All of that data is compiled in a single app, which also offers coaching to help drivers avoid dangerous situations like crossing the centre line or tailgating.
“We can keep vehicles on the road longer, lower maintenance costs and shorten service intervals by decreasing the amount of harsh wear and tear from driving. The great thing about this business is the returns on investment are so big and so fast,” says David Prusinski, executive vice-president of sales and marketing.
Fleet Complete charges between $20 and $40 per month for its tiered services. Given that a heavy-duty vehicle can burn nearly four litres of fuel if left idling for an hour, that cost could be recouped just by reminding a driver to turn off their vehicle.
Of the 600,000 units the company tracks and monitors all over the world, about 20% are something other than vehicles. The company has followed the location of everything from generators and chainsaws to porta-potties and cows. “If it is an object that has value and needs to be tracked from a maintenance-scheduling perspective, a usage perspective or a safety or theft perspective, we can track it and probably have at some point in time,” he says.
A portion of the growth in this sector is a direct result of a reduction in the price to produce and monitor tracking units. A device that once cost $200 to buy and $15 monthly to monitor now sells for $50 and can be tracked for $8 a month. “A couple of years ago, it just was not economically feasible to be able to sell a solution that would help because of the cost of batteries and chip sets,” says Prusinski. Part of the lower cost is mobility carriers launching narrow-band IoT networks in the past year, he says.
While his excitement for his firm’s current offerings is palpable, Prusinski gets even more enthusiastic when talking about where the technology is going in the years to come. “IoT solutions are going to start to break down to the component level of vehicles,” he says. Take, for example, the transmission in a 53-foot tractor-trailer, which costs $10,000 to $15,000. By tracking its wear and tear, the owner can better understand when the components will need service and avoid a roadside breakdown. In turn, that means dodging late delivery fines and spoilage fees, on top of the cost of repairing or replacing the transmission. “If you got it looked at ahead of time, maybe you could solve the problem for a few hundred dollars instead of thousands, all because you put a $150 IoT tracker on that transmission,” Prusinski says.
Just the right mix of data and fertilizer
by Kim Hart Macneill
Farmers today face ever-increasing challenges, ranging from the effects of climate change to global trade wars. Meanwhile, the demand for food is rising with the world’s population. Farmers are struggling to keep up—and keep their fields—in these demanding times. Decisive Farming in Calgary has created digital platforms to lighten the load.
Remi Schmaltz’s family has worked in agribusiness for four generations. Using the family firm, Schmaltz and his brother, Tasha, have been incubating ways to digitize a range of farm operations. In 2011, along with business partner Garth Donald, the pair took Decisive Farming independent. Today the platforms are used to monitor 40 different crop types across five million acres in North America.
OptimizeRX, the company’s flagship product, pairs graphic information system, or GIS, mapping with soil analysis to provide farmers with precise advice on where to place both seed and fertilizer. “It’s not about just cutting or saving on fertilizer—it’s about having the right amount,” says Schmaltz. That could mean using more fertilizer in better spots. “You could be spending more money, but you’re going to see that return at the end of the day.”
Decisive Farming’s cloud-based app, My Farm Manager, enables employees and service providers to manage data ranging from crop yields to equipment sensor statistics. For example, if a farmer changes their crop plan in the platform, it will update the information in their marketing strategy. “Having that information in real time and moving instantly between all the different partners and the farm is really critical today to making better decisions,” says Schmaltz.
As consumer demand for transparency about the origin and environmental impact of agriculture grows, food retailers have passed the responsibility of tracking that information onto farmers. “As an industry, we need to get ahead of the game when it comes to this digitization piece and having really good records, because in the future, those are just going to be table stakes,” says Schmaltz. “If you’re not there, you’re not going to be in the game.”