Artificial intelligence, Internet of Things, machine learning are buzzwords on everyone’s lips. Many of us are sick of hearing them. Must everything be smart and powered by AI these days? At least spare our beloved mountain bikes from it… right? We took a peek outside the box, looking to the automotive, aerospace, and IT industries for inspiration, and asked: what if…?
Picture this: you’re riding a trail for the first time, and loving it! Flowing berms, rock gardens, and then a drop – but you came in way too hot. You’re about to bottom out hard! But wait, your bike already knew what was coming and quickly increased the compression settings in the fork and shock, and also increased the tire pressures slightly – you land perfectly and carry on without skipping a beat. Impossible, you say? We beg to differ! It may not be available here and now, or not (yet) for mountain biking, but technology like this has already been implemented in cars, motorsports, and even aerospace. So let’s suspend disbelief for a moment. In our minds’ eyes, let’s hit the trail aboard an intelligent mountain bike and cast a critical glance at what the future of mountain biking may hold. We did just that, conjuring up visions, transferring existing technologies to mountain bikes, and also evaluated AI and Internet of Things technologies from a moral, ethical, ecological and economic perspective.
Glossary: Intelligent, smart, self-learning, automatic, or adaptive?
When we talk about smart or intelligent solutions today, we often mean the same thing: devices that make our everyday lives easier in certain ways. Case in point: smartphones. And as the name suggests, this is what you call smart. Thanks to its user-friendly interface, capability of connecting with a wide range of data sources, and the fact that it almost always has Internet access, the device is considered smart in the sense that it’s particularly useful. However, it’s far from being intelligent. This also applies to mountain bikes. Electronic suspension that can adapt its compression settings to suit the trail or riding style with the help of various sensors is neither intelligent nor self-learning. Rather, it’s smart and adaptive – one up from automatic. Ultimately, you must distinguish whether a system’s “intelligence” is based on predetermined algorithms, within the limits of which the smart system can actively adapt. Or whether a system actually learns from “experience” and develops new algorithms. Only then can we speak of artificial intelligence, in short: AI.
What if…? The AI already on the trails today.
Adaptive and responsive suspension, automatic shifting, ABS – various electronic technologies have long found their way onto both e-assist and analogue MTBs. But to what extent does the current generation of bikes already have something to do with AI? And what might that mean for the future of our sport? Sure, electronic suspension automatically locks out on the climbs, or opens the compression valve when things get rough. But is it intelligent? Or would that be capable of more?
The same applies to the electronic, automatic shifting, like Shimano and SRAM already offer – at least for eMTBs. But to call this smart or even intelligent would be more than presumptuous, and no better than the marketing hype spewed by its manufacturers. Yes, these drivetrains shift gears automatically and magically select the appropriate gear for pulling away while slowly rolling up to a traffic light, but they’re just pre-programmed algorithms. But disillusionment comes soon after hitting the trail. How is the drivetrain supposed to know that you’re about to hit a steep climb right after a fast descent, or that you would prefer having a high gear as you exit a tight berm so that you can get on the gas? That’s where our vision comes in… What if it did know?
What if your suspension knew what was coming?
So far, electronic suspension has simply been what it says on the box – electronic suspension. It’s designed to adapt the damping to the terrain, on the fly, always delivering the best suspension performance. Doing so requires a few acceleration, position and gyro sensors, an “intelligent” control unit, and a few batteries. And a sensor on the crankset tells the suspension whether or not you’re pedalling. That’s basically it. And even if the suspension adjustments happen within a few milliseconds, it’s got nothing to do with intelligence or anticipation. In principle, the system only reacts to the data it’s fed by the sensors, and adjusts the compression valves accordingly. Therefore, it’s merely reacting, albeit very quickly.
However, in our vision, smart suspension would draw on a lot more data and intelligent suspension would also act with foresight and adapt. The only limits of what this could mean are the limits of your imagination. Even now, with existing technology, electronic suspension could detect if you’re about to take a dive, compressing the suspension until it bottoms out to make the bike more compact, which would make it easier for you to eject safely. Or prevent you from getting knocked on the back of the head by the bike as it cartwheels behind you and eventually catches up. With information on the temperature and the rider’s weight, the suspension could adapt accordingly – especially after a big lunch break. The required sensors could easily be integrated into current bikes. What if the suspension control unit had access to real-time data from the Internet, which would be easy in view of smartphones and modern GPS bike computers. Your bike could know what’s coming up on the trail ahead and adjust the setup and geometry accordingly. Keyword: hive mind. The data collected from other riders that have gone before you is the best foundation for truly smart, predictive suspension. The likes of Strava receive an insane amount of data every day, which still remains completely unused or, at best, is simply used for statistics and bragging rights. Combined with live weather data and the current state of the trail, the suspension would be prepared perfectly for all eventualities. Nothing but fantasy? We will see.
What if your drivetrain knew which gear you need at every moment on the trail?
Automatic shifting – if you ask the two big drivetrain players, they’ve already answered this question. While this may be largely true for roads and cycle paths, they can’t quite deliver on the trails, as we found out in our reviews of the SRAM Eagle Powertrain system and the Shimano XT Di2 drivetrain with AUTO SHIFT. Unsurprisingly, our beloved sport thrives on variety and unforeseen twists and turns. Current electronic shifting is far less partial to surprises because it lacks information. First of all, drivetrain control units would need more data than just speed, pedalling cadence, and watts. Like data on the terrain and current gradient, supplied by the electronic suspension. Of course, it would be even cooler if the drivetrain also included your current physical state and riding style. With a smartwatch and a small memory card in the bike that shouldn’t be a problem. Neither should taking into account the experience, or rather data, of thousands of riders who have ridden the same trail before you. Which brings us back to the hive mind, and the cloud, and to data privacy and protection… But more on that later. Shimano strive to make their DEORE XT Di2 LINKGLIDE drivetrain capable of learning. To quote from the review conducted by our sister magazine E-MOUNTAINBIKE: “…there’s a learning mode in the pipeline, in which the shifting algorithm can be trained to react to different riding situations. Let’s say, if you often override the system manually to shift into a lighter gear when negotiating a steep climb with a higher cadence, the drivetrain could learn to recognise the pattern and shift into a lighter gear automatically.” That wouldn’t require access to the Internet or the cloud – excellent. However, that looks to be about simple adaptation, and not true, AI-powered self-learning.
What if your eMTB always delivered the power you really need?
Electronic components always need electricity. Not much, but they still require a battery. That’s where eMTBs can shine because they’ve always got a battery on board anyway, and with relatively large capacities, too. They can supply all kinds of electronic gadgets with power without the need for additional batteries. But first things first: a smart drive unit would be a good start. A motor that always delivers the right amount of power depending on the terrain, while taking your current fitness level into account. This data can easily be obtained from a smartwatch or chest strap. The brains at Specialized have shown what’s possible with their app, at least to some degree – and that was a few years back. Among other things, the Specialized app allows you to specify the length of the route and desired minimum residual charge when you arrive at your destination. Based on this, the system then automatically selects the appropriate support mode. A third-party app even allowed you to specify the rider’s power output and heart rate, so that the motor only provided enough support to keep the rider at the desired level. That’s until Specialized blocked Bluetooth access for third parties. With the appropriate sensors and data on the weather, temperature, and topography of the planned route, the system could ensure that the battery lasts until the end, but the rider doesn’t have to exert more effort than necessary.
But we’re just getting started, especially when it comes to motors and drivetrains. Just think of the automotive sector, where things like traction control have long been commonplace. On technical climbs, the motor could adapt its output to the terrain at lightning speed and prevent the rear wheel from spinning out, like when riding over wet roots and rocky ledges. In combination with a smart sustained response, you could tackle climbs aboard your eMTB that would otherwise be impossible even on foot.
What if your bike was the best trail scout of all?
Riding in unfamiliar terrain is incredibly exciting and often just as frustrating. Discovering new worlds, shredding unknown trails, and simply making new experiences aboard your bike – this is the essence of our sport. But this often comes with unpleasant surprises: closed trails, foot paths where bikes are prohibited, impassable terrain. Sure, with the likes of Komoot, you’re not completely in the dark when exploring many trails that are new to you. Routes can be planned, and the best trails are just a finger tap away… And yet, things hardly ever go according to plan. With access to the internet via your smartphone or GPS device, exploring new terrain can be much less stressful. And thanks to the hive mind, even supposed insider tips didn’t go unridden. We could even imagine live tips like, “Skip the next trail. It will hardly be ridable after three days of continuous rain.”
What if… Almost anything is conceivable, but does it make sense?
We could continue with these visions and wishes almost indefinitely – there are hardly any limits to our imaginations. Like dropper posts, for example, which certainly still have a lot of potential for smart features. They could drop automatically the moment you get out of the saddle while approaching a steep descent. Ideally, your dropper will know that there’s a steep section coming up before you even see it. There are a lot other features we can think of for the drivetrain, brakes, and suspension that have long been proven in the automotive sector, like wear indicators that could help you know when to replace components like the chain or brake pads well in advance, so they don’t let you down on the trail. Speaking of brakes, you could have a brake fade or overheat warning light on the reservoir. Or stretch marks that show up in the frame before it breaks or cracks… As you can see, the list is endless. But which of them make sense, and what’s the limit of what’s possible, affordable, and reasonable?
The intelligent MTB – More to go wrong and get sabotaged?
So, we’ve established that there are countless ways to make our beloved bikes smart and even intelligent. And where there is an obvious benefit, it certainly makes sense to use modern technologies like AI or the hive mind. Adaptive suspension, air pressure control, automatic shifting, ABS systems – all these are features that can make mountain biking easier and safer. All of which have already stood the test of time in the automotive sector. And all of which have made cars and motorcycles increasingly expensive. Unless you’ve been living under a rock, you will also have seen this trend in the MTB sector – even discounting the perfect storm that was the pandemic and subsequent supply chain issues. Nevertheless, new technologies always drive up the price, and often it’s only the early adopters who are happy to pay for it. However, regardless of the cost, you must ask yourself about the downsides of smart or intelligent MTBs?
There are the obvious things like the need for batteries wherever you’ve got electronics. A bike with wireless shifting, electronic suspension, and a wireless dropper often has 5 separate batteries, and that’s not counting the button cell batteries in the remote controls. They’ve got to be charged, have a limited service life, and can get lost. And the electricity consumption isn’t negligible. Sure, shifting a gear consumes almost no electricity, like snapping a photo with your smartphone. But, as always, it adds up: tens of billions of photos are taken, shared, and backed up every day. According to a study by the Carbon Trust, Instagram alone consumes about 170 GWh of electricity per year, which is roughly equivalent to the energy requirements of a medium-sized city. Instagram alone!
So, is the additional electricity demand justified? And what about the millions of batteries needed to power the electronic components? Of course, like with EVs, environmental, moral, and ethical aspects play a role here, too. Even if you only include the mining of raw materials such as lithium, tungsten and cobalt in the equation. Furthermore, there are other factors besides that have nothing to do with costs, energy, and moral responsibility. Over-engineering, for example. Do we want bikes that are overloaded with technology and features that we need to wrap our heads around? Adding complexity and things that can break. Or do we want uncomplicated, reliable bikes that we can ride whenever we want? Bikes that can take a beating and aren’t overly sensitive to the elements and external influences.
Finally, smart and, above all, connected devices are always at risk of sabotage, though it’s more of a theoretical issue. Just imagine someone standing by the final climb of a race stage with a 2.4 GHz jammer, making all wireless shifters fail… or even worse: someone could access your bike’s control unit from their smartphone and lock your brakes.
The intelligent MTB – What happens to your data?
The mention of AI and the cloud is sure to make many of you prick your ears. Because wherever there’s big data involved, there is always the risk of abuse. If it’s just about gathering information about routes or trail conditions, that doesn’t seem so bad. But maybe you don’t want your movement profile published on the internet for everyone to see, advertising where you live and ride. And very sensitive data like your vital signs certainly shouldn’t land in the wrong hands. Just imagine if employers had access to details about the chronic conditions of their employees… Or health insurance companies knew all about your leisure activities and the number of post-ride beers you’ve consumed. The topic of data privacy and protection is more important and also more complex than ever in our data mad world. So, if our beloved mountain bikes keep getting smarter and more connected, then brands must be able to guarantee that our data is in safe hands. How can they? Good question!
Intelligent MTBs? Yes, but in moderation and within reason
Our vision of the intelligent MTB sounds both nerdy and exciting. However, with adaptive suspension and automatic shifting, we’re already seeing the first attempts at making MTBs smart. But only time will tell what the future holds. Almost anything is possible, but is it also feasible? Ultimately, the added value, price and, above all, customers will decide how far the industry can go with connectivity and artificial intelligence on MTBs. We’re fine with it as long as it doesn’t compromise our beloved sport!
Did you enjoy this article? If so, we would be stoked if you decide to support us with a monthly contribution. By becoming a supporter of ENDURO, you will help secure a sustainable future for high-quality mountain bike journalism. Click here to learn more.
Words & Photos: Patrick Gruber, Peter Walker
