19-01-2021 | | By Robin Mitchell
While fully autonomous cars are yet to be developed, many cars already integrate various autonomous features. However, why are some modern cars struggling with cattle grids, and how does automatic breaking work?
In an ideal world, cars would not require breaks, traffic lights would be irrelevant, and indicators would serve no purpose. Drivers could apply the accelerator when needed and then roll to their destination while co-ordinating their route with nearby drivers.
Of course, this is the real world, and anything can happen at any time meaning that drivers must be ready to perform an emergency stop, use indicators to show where they are going, and be alert for other drivers who may not be as good as a driver. From dangerous drivers to falling trees, no driver can predict what will happen, and in some of these cases, drivers cannot react fast enough with the result being disastrous.
One recent development in car technology is automatic emergency braking which can automatically perform an emergency break upon detecting an obstacle on the road. Before we understand how this system works, we must first understand what breaking assist is and how it is not the same.
Braking assist is a mechanism that detects how the driver uses the brake and applies extra force to the break line if it detects that the driver wants to perform an emergency brake. For example, the driver may suddenly apply substantial force to the break, but not apply it fully. In this case, the brake assist would detect this and add the extra force needed to apply the breaks fully.
Automatic emergency braking, however, will apply the breaks if an object is detected regardless if the driver has used the breaks or not. Depending on the specific technology, automatic emergency breaking utilises either RADAR or SONAR to determine the distance between the car and obstacles ahead. If a sudden drop in distance is recorded, the onboard computer can engage the breaks to prevent the car from colliding with the obstacle. Next-generation systems may utilise cameras and LiDAR as these technologies will become critical in future autonomous driving systems.
What are cattle grids?
Unknown to many urban drivers, countryside roads are not as forgiving as city roads with gravel roads, plenty of potholes, small narrow lanes overgrown with hedges, and blind junctions. In areas that deal with animals, including cows, horses, and deer, roads can also contain cattle grids.
Cattle grids are hollow metal grids embedded into a road that looks something similar to a bridge. However, instead of allowing things to cross, cattle grids prevent livestock movement as most animals have feet (or hooves), with a small surface area. Thus, any such animal that attempts to cross will find that their feet slip into the gaps.
Cattle grids are important in areas where animals may be able to roam freely, easily find ways of wondering, and control animal movements. Anyone who has driven over such a grid knows the horrendous shaking and sound, and those with luxury cars would be forgiven for turning around and not even attempting to cross one.
As strange as it may seem, some modern cars with automatic emergency braking systems have been performing emergency stops when approaching cattle grids despite the grid not being an obstacle. Somerset council, who have been most affected, have noticed that many cars have been so badly affected that the vehicles are skidding off the road into ditches in an attempt to avoid the cattle grids.
To combat the situation, Somerset council are spending £70,000 to make adjustments to the grids so that such cars will not perform unneeded emergency stops. But what is causing these vehicles to stop, and what does this mean for autonomous systems?
According to the council, the cattle grids need to be level, and country roads are not known for their levelness or smooth gradients, and so such grids had sudden gradients. These sudden gradients to a car may appear to be a wall and can trigger the automatic emergency brake.
To a normal driver, a cattle grid with a steep gradient is obviously not an obstacle or wall, but to the automatic braking system, it was. Therefore, the next generation of automatic emergency braking can be programmed to recognise such obstacles better, but is that really the best solution?
The biggest hurdle for autonomous systems in cars is to handle any situation in an identical way that a human driver would. Cattle grids are not the only unknown factor in driving, some countries have big speed bumps, while others may use a different pattern in the road not commonly seen to indicate a pedestrian crossing.
If autonomous vehicles are to become a reality, they need to be able to properly understand the environment around them and make decisions around obstacles in the road. Is that anomaly up ahead really an obstacle? Is it safe to perform an emergency stop? Is there anyone behind me?