Self Driving Cars

Self Driving Cars

Photo by Shubham Dhage / Unsplash

Autonomous Vehicles. Very realistic & coming to a dealer near you, much, much sooner than you'd think.

Prior to taking a real deep dive into the topic, I had mentally likened autonomous vehicles to the concept of flying cars; highly speculated but never brought to fruition. Automakers are yet to deliver on this promise to enthusiasts like myself.

Side Note - Flying cars

I actually heard recently that arguably, we already have flying cars... they're called helicopters (I think it was Neil deGrasse)- He went on to say that there will never be flying cars in the sense that we have all imagined. Main reason being, the shear amount of noise produced in generating enough uplift to raise a car. Drones are noisy enough, could you imagine the sound of a human sized drone? Other pitfalls of the technology include stark inefficiency and piles of associated dangers to be solved. Squadrons of engineers are struggling with the balance between making a vehicle that is robust enough to withstand the bumps and knocks of the road but yet light enough and aerodynamic enough to enable safe take off, flight and landing.

Flying cars would significantly reduce congestion on the roads as it offloads traffic into another dimension: the 3rd dimension. Cars flying at different altitudes seems conducive to avoiding congestion but this only puts the image, in my mind, of a mixture of incompetent drivers and overaggressive imbeciles in the air. A recipe for Skyfall.

There are many prototypes out there but they are all a long way from providing solutions for the aforementioned concerns, I guess we will just have to wait and see what the future holds.

Automated Cars

It seems that efforts have since been siphoned to advancing self driving car technologies. Doubling down and optimising 2 dimensional travel. Self driving cars would also reduce the traffic on the road as there would be an internet of cars communicating with one another (& the actual road), as well as traffic lights, signs and all other associated technology. Therefore, allowing things to run seamlessly even at speeds of up to 120mph! The following video shows the operation of robots on a factory floor, providing premise for the viability of autonomous, communicating vehicles. "If it can be it can be made to work on a small level, it can be scaled up."

But what's the hold up?

As you can imagine, anything that involves or could endanger human life falls to heavy scrutiny, meaning every aspect must be considered meticulously.

We will start off with the most recent updates and identify the gold standard for the technology, pros & cons and then explore the likelihood & possible dates of seeing a fully pledged driverless vehicle on our roads.

There are many companies working in this space with the front runners being: TESLA, HONDA, TOYOTA, WAYMO(of Google) & CRUISE. They have all put products out there that show the technology is functional and relatively safe. However, there are still hiccups in the tech which are being ironed out. We draw closer to an appropriate solution year by year.

Speculation still remains in the news of when these cars will be deployed for public utilization. Articles quoting "By the end of the year", "Next year" and "Tomorrow" flood the internet, but with 3 more days left of the year, we are hardly convinced. Changes in legislation however, provide some hope of steps in the right direction. The Department for Transport (DfT) announced that it would allow hands free driving in vehicles with lane keeping technology and also intends to add to the code that, drivers may be allowed to watch TV programmes on built-in screens whilst using self driving cars, but use of mobile phones will still be prohibited (where is the logic??). These changes by the government will go a long way in supporting the adoption of automated cars.

At the moment, Tesla offers software packages that grant the cars the ability to: accelerate, brake, steer, change lanes and park, all without human intervention. However, CEO Elon Musk says the vehicle is not yet able to satisfy authorities of its autonomy and that the technology will still require human over sight, where drivers will rarely have to touch the controls but will need to remain fully attentive at all times.

Waymo and Uber are pushing forward in an effort to provide driverless cab services. In Las Vegas, Uber is already offering the option to book autonomous "robo-taxis" through a third party company called Motional. However, there are trained "vehicle operators" behind the wheel to monitor the technology and provide support if needs be. Waymo also has test driverless vehicles in Arizona in a tightly controlled 50 square mile area.

Levels of Self Driving Cars

There are 6 stages of autonomy in vehicles and they are described as follows:

Sourced from Openeo.co.uk

Currently, vehicles are hovering between levels 2-4 with level 5 being the ultimate goal. No road car has yet passed the level 4 ranking for safe commercial road driving.

How Do Self Driving Cars Work?

There are a myriad of cameras, sensors and computing technology that must all work in unison to allow proper functioning of the vehicles. The cameras/sensors allow for seeing and gathering information of the surrounding environment and a strong computing system processes it all in real time.

The sensors and cameras located on the cars include:

  • 3D video cameras placed all around the car to allow a full 360 degree visual of the surroundings. These produce a highly detailed and realistic image of the environment. These cameras are heavily affected by poor weather however.
  • Radar sensors send out and receive radio waves that aid in figuring out relative distances and speeds of the surrounding objects in real time. Both short range (24 GHz) and long range (77 GHz) radar sensors are used to allow full functionality of the car's features. Radars are not affected by weather conditions.
  • Lidar sensors, short for "Light Detection and Ranging" work in a similar way to radars, with the only difference being the use of lasers as opposed to radio waves. Lasers allow for 360 degree 3D mappings of the space and objects around the car as opposed to the 2D horizontal beams in many radar systems. But they are significantly more expensive due to being made of rare earth metals.

The computing technologies pulling it all together are artificial neural networks (AI), Machine Learning (ML) and specific intelligent algorithms. These work together to aid in object identification, depth perception and route planning. Blockchain technology will also assist in creating this IoT.

Types of Self Driving Vehicles

There are 4 main categories and they differ based on whether they travel mainly on motorways or city roads, and whether humans can get in them or not.

The 4 cases are: Commercial Vehicles owned by all (1), Long Distance Transportation Trucks (2), Robo-Taxis (3) and Food Delivery Services (4). These can all be visualised in a quadrant like so:

I'm sure they will all work abreast but they will probably be rolled out at different times. Door dash already has the delivery robots deployed, we will probably see the robo-taxis next, followed by either the haulage trucks or the commercial cars; hard to say which will be first.

Pros:

  1. The main positive for the case of driverless cars is the removal of human errors. Machines do not tire and their attentions never wander. Their cognitions are not altered by substances like drugs and alcohol. They've no messages or calls to answer and thus can just do what they are supposed to do, 100% of the time, 24/7. Humans will no longer be put behind the wheel of a large chunk of metal driving at 70mph whilst inebriated and checking Waze in their other hand. This will result in, safer roads, significantly less motor vehicle accidents and in turn less deaths.

    According to the World Health Organisation (WHO), 94% of all car accidents are due to human failures. Over a million people a year die from car accidents resulting from these driver errors. Due to the algorithmic nature of the machines, there would be no excessive speeding, no aggressive over takings, no state-wide police chases. IN THEORY, although it is never the case in practice, 94% of all accident should then be cut out and thus approximately 94% of lives should be saved per year (940,000 lives). That would be a remarkable difference!
  2. There would be far less traffic with the increasing spread of autonomous cars. Have you ever been in traffic and wondered what is causing this all? Bar bad weather conditions, it is again down to human incompetence. Accidents, humans slowing down to look at said accidents and mainly the "Shock-Wave Effect": Shockwave traffic jams recreated for first time - YouTube
    With autonomous vehicles, if the speed limit is 30mph, it would be set and there would be minor deviations from this, no unnecessary braking. This ties into the Internet of Things concept.
  3. Internet of Things: there would be a highly interactive network of machines connected and communicating through blockchain technology. This would allow for better and more efficient travel. A vehicle would know how many other vehicles are on the roads and at what times, giving for optimal route planning to avoid excess congestion and far more accurate ETAs. Cars will gather information from all traffic lights and signs on the routes, be better updated about road closures and accidents and the technology will be continuously updated. All these stresses will be lifted from the rider's mind.
  4. You could take your car with you on a night out!
    You could enter your car, it'd drop you right outside the venue or location, it would find a spot and park itself, you could summon it once you've challenged your liver enough to your liking and hop it sheepishly into the comfort of your own car, heading right for your doorstep. All legally, no concerns of breathalysers.
  5. Another notable benefit of these technologies would be the access that physically and mentally impaired passengers would now have to driving. A range of individuals with medical conditions like blindness for example, would now be able to own their own cars, and live more "normal", self sufficient lives.

Cons:

  1. Lack of control: as safe as the tech may be proven to be, there is still the underlying risk that the vehicles may go rogue. I'm reminded here of I, Robot or Westworld. What happens if robots become sentient? What if robots become "makist" and get into ramming wars with cars of other manufacturers? So many things can go south and if we lose the ability to switch off these robots, all hell will break loose.
  2. Who would take responsibility for accidents? The owner of the car? The insurance companies? The car manufacturer? How do you assess who is at fault? What happens when an human operated vehicle runs into an autonomous vehicle? There are a cocktail of unfavourable scenarios that may arise in this new environment, interpreting and judging them will be a whole task in itself. The DfT stated earlier this year that insurance companies will be liable for claims rather than individuals in the event of an accident. What if it were a manufacturing error?
  3. Driverless cars take the fun out of driving. What happens when all cars are autonomous? Will individuals who actually find enjoyment in driving be forced off the roads? Will the future generations be required to get licenses or will a rich kid as young as 12 be able to be a car owner?
  4. The automated cars may potentially be over pre-emptive. This is the polar opposite of their efficiency. What is too many cautious measures are instilled in the programs causing the tech to cause more traffic than it reduces?
  5. Technical issues: (bunching these together to save time) could a system be hacked? Human-Robot Interactions? Poor weather conditions?  Shadows? Misidentification?

I'm sure you're getting the vibe by now, there are so many unanswered questions hanging in the air and the truth really is, we will all have to just wait and see and manufacturers will have to be reactive and adapt swiftly, rolling out new updates regularly.

Probability of Materialising:

The question here is not if, but when. The real bottleneck is solving real world AI. Once this is achieved, the early adopters will begin to grab their toys!

Driverless cars will probably get the green light in the next few years but there will likely be a long period of transition before they become ubiquitous. Predictions have been made that no car on the road will be human operated in 25 years time, in the year 2050.

As autonomous cars begin to be rolled out onto our roads, initially there will be a mixture of human operated vehicles and computer controlled vehicles and it will be interesting to see how these interactions play out. How will driverless cars respond to driver errors and the unpredictable human nature? More than likely, accidents will still occur.

In the end, all cars will be autonomous and will communicate efficiently between themselves and this is when the significant decline in motor accidents will be realised. If a self driving car wants to go left, it would tell the other self driving cars around it "Mind out, I'm turning left" and they would simply move in a way to accommodate this.

Widespread mass adoption will require two things: legislative approval & global confidence in the technology. The latter will take place when we start to see other humans get into driverless cars, use them and most importantly, get out safe and unharmed. The fear will subside and it would quickly become an integral part of our lives as it usually does with new technology. Take the case of elevators as an example. Initially, they were viewed as a death trap. An elevator operator was put in place in order to exhibit some sort of human control. Slowly the narrative changed and now we all use them without the fear of an impending falling death - well, that is, until the elevator jolts.

In the interim, whilst mass adoption is still being achieved, some suggest a separate lane for these cars where they could operate optimally, parallel to all human activity.

Robots are not inherently bad or evil, they are just tools. If we get robotics and automation right, we will free up a lot of time to carry out our "humanness" and express our creativity.


Here are some other topics I found whilst plunging into the rabbit hole that you may also find interesting:

  • Flying Cars
  • Hyperloop Trains
  • Lithium-Silicon Batteries
  • Wave Renewable Energy

Peter

London