The history of “engineering inventions” is almost as long as human history. Afterall every tool, from the very first flintstones, is an “engineering creation” and this is what sets us apart. Over the past few centuries the scientific breakthroughs and the engineering invention rate grew exponentially, where some inventions (e.g. transistor, wireless communication, internet etc.) opened a whole new space for “engineering creation”.
From an early age I was fascinated by “engineering inventions” of all kinds, and their history. Aside from the excitement of learning something new, the frequent question that came to mind was “how did they get to that idea?”. I guess that this puzzlement is common to creativity in every field. Hearing a “mind blowing” musical piece for the first time, something you haven’t heard before (did anyone say “PInk floyd — The Dark Side of the moon” ) — all composed of the same notes being used by others for centuries and yet… . Or reading a book that is taking that old medium to a different dimension , or even a TV show that is inventing a whole new concept (“The Singer in the Mask”?).
For years I used to think , as many others, that inventions are mostly a product of rare moments of inspiration and enlightenment, ”Eureka” moments (following the famous story about Archimedes in the bath). And then , during my BSc studies I was exposed to the novel concept of “systematic invention process”. You might have encountered it by the name TRIZ (by the Russian origin of the method, 1946) or TIPS ( Theory of Inventive Problem Solving in English) or other versions that have been developed over the years. The idea is that there could be a method to “invent”, to come up with a creative solution to a problem. The method revolves around several techniques that could be applied to different challenges. The methods differ in the process, the definition of the problem and the “rules” for “solutions search”; but they all take you through non-trivial spaces to find new solutions to old problems (hence “invent”). A simplification of TRIZ that was developed in the 90s is called Systematic Inventive Thinking (SIT) and it lists the following thinking tools (credit to https://bold.group/):
Removing essential components from a product, process, or business model (examples: iPod Shuffle (removed the screen), Dyson hand carried Vacuum Cleaner).
Adding existing components to a system and giving them a twist (examples: Gillette razors back-side razor, double-glass window)
Assigning new tasks to existing resources (examples: mobile phone camera, pedometer watch)
Changing variables in a running system (examples: airline booking classes, happy hour)
Dividing in space and time (examples: deferred payment credit, remote control)
This is all very nice but how does it relate to traffic monitoring?
To answer this question we should ask ourselves what is the problem we are trying to solve and refer to the current solutions. Then I would show how the above method is practically applied today by Valerann to propose a paradigm shift in traffic monitoring technology.
In general, there are two aspirations for ideal traffic monitoring: 1) To be able to observe everything that relates to normal and abnormal traffic regardless of weather or road conditions. 2) To serve action-oriented monitoring (e.g don’t overwhelm the operator with non-actionable data).
The popular current solutions (magnetic loops, road side cameras, road side radars) are far from satisfying those needs. In practice they give very partial coverage of the road , and deliver very raw, mostly non-actionable data.
How would SIT approach it? Can we look into one of the existing solutions and modify it , applying the above techniques, to come up with a different solution that overcomes the mentioned challenges?
Let’s look at the camera for instance: what can we change such that we’ll get full road coverage by cameras? In all whether , day or night? Let’s apply the ”multiplication” technique: Use many short distance cameras instead of a few long distance cameras. This would give a full road coverage but is not practical in terms of cost (for the device, it’s installation and video stream communication and processing fees) and is still subjected to limited performance under heavy rain/snow/sun glare. So what about applying the “subtraction” technique? Get the CMOS (the semiconductor light sensors) out of the camera (…yes, I know it sounds like a contradiction , but sometimes this is what the subtraction technique is about) and replace it with another sensor? Cheaper, weather-agnostic? E.g. magnetic field sensor? Will it do the work? Would it be able to give full road coverage, in all weather, at an affordable price? Will it meet the required performance? It isn’t trivial to overcome the new challenges but this is exactly the nature of an “engineer breakthrough”: it trades a known set of hard-to-overcome challenges with a new set of challenges. The evolution is no longer linear and incremental but done at a different, new, dimension.
The IoT sensing world is all about that: replacing legacy methods that typically involved sporadic and cumbersome sensing with on-going, affordable , full-coverage sensing by many relatively small and cheap sensors. The fact that the underlying sensors are usually cheaper and less sophisticated is being compensated by either being closer to the sensed object or by post processing algorithms that are applied to each sensor and to a group of sensors as whole. Obviously it’s advantageous also in terms of reliability due to the inherent redundancy, which translates to lower solution cost too.
The progress in processors, batteries, and sensors pushed by the smartphone industry and the evolution of the internet and big data cloud infrastructure positioned the IoT sensing today to drive a paradigm shift in many cases of monitoring technology. Examples range from agriculture (see CropX, https://www.cropx.com/, for example) — where sporadic soil samples were replaced with full ongoing soil and crop monitoring, to manufacturing lines were the manufactured object (being a mechanical bearing or electronic circuit board etc.) is not just being tested to meet the spec when it exits the line but is monitored along many steps during the process (and obviously the machines involved are continuously being monitored to meet their own specification, see Augury, https://www.augury.com/, for example).
Valerann is taking advantage of IoT smart sensing to push a paradigm shift in traffic monitoring and offer full coverage, at any weather, as well as apply advanced algorithms to focus the road operator’s attention to road incidents.
Valerann’s Smart Road Sensor system is based on an in-house development of smart sensing devices that are fit to endure and function from within the asphalt and provide direct, robust and cheap sensing. The sensors are deployed along and across the road, spaced 20–50 meters apart from each other. It takes advantage of the IoT principals where all units are the same, its deployment is simple and doesn’t require any specific tuning to location or orientation and is maintenance free and uses a very efficient communication channel. It overcomes the traditional limitations of road side cameras/ radars and harnesses the power of big data algorithms to re-create a reliable traffic picture and observe abnormalities.
Imagine a car is driving 60mph at night and in heavy rain. Down the road a cardboard box is laying in the centre of the lane. The driver is approaching it fast but manages to zig zag and avoid it. The traffic is very light and no harm was caused. It came by a surprise though and it was very close to ending differently. Many drivers experienced such a situation. It isn’t that rare. It could be very helpful to be notified of it in advance and to notify others, but it is dark and raining heavily and the driver’s attention should be focused on the road, so they do not call 911 and definitely does not initiate a Waze alert. So what happens next? Usually not much…until something happens (most often the box is hit to the roadside , but every once in a while the ending is much worse). Sometimes, if it is in a spot with a decent road-side-light and with a nearby camera pointing at it, and the camera is not obscured by the heavy rain , and the camera isn’t getting blinded by the other lights smeared by the rain, and there is an operator in the control centre that is watching this specific camera footage and is alert enough to observe the box or the abnormal manoeuvre of the car (note how many “if”s over here), they would send a road patrol / set the right overhead message in the road signage and prevent the escalation of the situation.
Now imagine it again, just this time the road is equipped with built-in sensors . No , they can’t see the shape or color of the cars that are driving, but it can clearly observe, directly and in any weather, the driving abnormalities, and call for the drivers attention (via the control center and overhead signage) — what a different ending. It is a dramatically different take on road traffic monitoring — and this paradigm shift is driven by IoT sensing.
In practice, sometimes it makes sense to fuse different kinds of sensors and to make use of existing assets, but augmenting existing systems with Smart IoT Traffic sensors paves the way for traffic monitoring at a totally new level of accuracy and reliability.
About the Author: Avi Tel-Or is Valerann’s VP R&D where he manages Valerann’s technology team.
At Valerann we fuse novel IoT approaches with the advantages of big data, and machine learning. Leveraging these techniques we are able to reach unique granular real-time insights that improve traffic safety, toll revenues, and enable roads to support connected and autonomous mobility. To learn more visit us at www.valerann.com