As far as technological showcases go, there’s nothing like bombing them to bits with an absolute flood of tech – there’s certainly enough of that in the BMW ConnectedDrive arsenal, covering a staggering number of topics. Our earlier story focused on the communications, infotainment and personalisation aspects, already impressive in its breadth and scope, and now, it’s the turn of the driver assistance-related systems, many of which merge into the Efficient Dynamics domain, and these are no less notable in both amplitude and presence.
Full story after the jump.
Active assistance for parking
Park Distance Control, no stranger to all, is moving forward, and actively. A new, extended version of the parking aid called Active PDC, which uses a three-stage intervention strategy (based on speed limitation, appropriate adaptive braking and braking to a complete stop) is the next step in offering greater protection from parking bumps and scrapes.
Active PDC offers the familiar visual and auditory feedback regarding the remaining clearance in front of and behind the vehicle, though for improved detection of the vehicle’s surroundings, the four ultrasound sensors at the front and rear of the vehicle are assisted by the laterally positioned sensors of the Park Assist system. During the parking manoeuvre, Active PDC limits the speed of the vehicle to 5 kph, and as an obstacle appears in the predetermined sensing range, the system reduces the speed of the vehicle down to 1 kph as the gap decreases.
If there’s a risk of hitting the object, the system initiates abrupt braking of the vehicle to a complete stop, taking into account the vehicle’s direction of travel and the steering angle. The system also primes the brake system in advance to further reduce response times – by braking or applying the accelerator, the driver can override or reinforce the function at any time without deactivating the system.
Park by remote
Not too confident of parking your car in a garage? Get Remote Controlled Parking to do it for you. The system can perform an entire parking manoeuvre – in this case forward perpendicular, or garage parking if you will – sans fuss, and you don’t even have to be sitting inside the vehicle. First presented by the company in 2006, the system has been improved on and refined – the new prototype of this system is capable of parking in any garage, even an unfamiliar one, without the aid of a reflector.
A driver planning to park in a narrow garage or a confined parking space simply gets out of the vehicle and activates the automatic parking manoeuvre using the vehicle key – RPC takes control of the acceleration, braking and steering functions and manoeuvres the vehicle autonomously into the narrow garage, or out of it again. The parking manoeuvre is activated by pressing a specific sequence of buttons within a set time limit, and the driver must remain in the immediate vicinity of the vehicle throughout. Activation of the system simultaneously locks the vehicle, switches on the parking aid and the dipped headlamps and folds in the exterior mirrors.
The previous prototype featured a camera and reflector, but the current one uses the vehicle’s existing sensor systems – guided by the six ultrasound sensors of PDC and Park Assist systems, the vehicle slowly moves forward into the garage at around 2 kph, aligning itself centrally between and parallel to the walls on either side. Necessary steering corrections are performed by the power steering system motor, and the control unit also works the engine Auto Start-Stop function, the gear selector and brake system. If there’s an obstruction, the sensors order an automatic stop and the hazard blinkers come on to signal that the operation has had to be aborted. If it’s all clear, the system completes the parking manoeuvre, brakes to a standstill and changes to selector position P.
Safety’s the big thing – whether going in or out of a garage, the last button in the sequence must remain held; if the button is released, the car stops immediately. For the exit procedure, if the driver’s door is not opened within a given time limit, the system automatically shuts off the engine and locks the vehicle. There’s also a distance restriction in which the vehicle can travel autonomously – in the case of the BMW 5 Series prototype, a maximum of seven metres. And with the law in many countries making it illegal to start the engine from outside the vehicle or to move the vehicle if the driver isn’t at the wheel, this one needs more than just tech to get going. Seeing it in action is quite the adventure.
Improved pedestrian protection, with the help of cameras
While cars have become a safer place for occupants, the vulnerability of other road users, namely pedestrians, hasn’t changed. Hence the need to go beyond passive protection, and preempt things before they get nasty. Pedestrian recognition with warning was added to the Night Vision system three years ago to help combat early morning and late night mishaps with pedestrians, and now BMW is developing a more proactive pedestrian protection system.
Here, pedestrians are detected using a camera. The camera-based pedestrian protection system is based on a detection and warning algorithm which detects potential pedestrian accidents using a sequence of camera images in combination with vehicle-related data. BMW isn’t new on this, of course – Volvo is the first off the block commercially with its Pedestrian Detection system as seen in its new S60; that one is also camera-based and works in tandem with auto-braking, operable at speeds up to 35 kph.
The system here is essentially similar. First comes the acute warning stage, which as with the Night Vision system, warns the driver of potential risks by means of visual and auditory signals – usually, the pedestrian will still be far enough away from the vehicle for the driver to be able to avoid a collision by taking evasive action or by braking. In parallel with this warning, the braking system is primed so that it will be able to deliver more braking power more quickly.
If it’s no longer possible for the driver to react in such a way as to avoid an accident, the system moves on to automatic braking. The driver can reinforce the system’s intervention by braking, but can also override and cancel the automatic emergency braking intervention by steering or accelerating. The automatic emergency braking, lasting no more than 600 milliseconds at most, takes place as late as possible, in order to prevent unnecessary activation.
It’s called AHB, in short
On the vehicle-vehicle front, Active Hazard Braking takes things further; at speeds of 80-130 kph, the research prototype uses autonomous braking to avoid an accident. In the current prototype, laser scanners and radar sensors pick up vehicles and obstacles up to around 160m in front of the car and up to 20m on either side. Additional radar sensors monitor the area up to 150m behind the car.
To ensure that situations are recorded as accurately as possible (to prevent misreading of say, a situation where the driver is merely preparing to overtake another vehicle), the sensors not only record the distance and speed to the vehicle in front, they also determine overlapping, acceleration differences and the time in hand in relation to all relevant vehicles within the detection range. They also monitor the entire area around the car, including the crash barriers and any constructions lining the road to check whether an evasive manoeuvre would be an alternative course of action
Again, at the first hint of impending trouble, the car first alerts the driver to the hazardous situation with a warning signal, flashed up on the Head-Up Display but also delivered in audible and tactile form – if there’s no reaction and an effective evasive manoeuvre cannot be had, Active Hazard Braking intervenes to slow the vehicle automatically, with the degree of stopping power graded as the situation demands to the point of emergency braking. It’s this anticipatory strategy that gives AHB a capability far beyond that of the emergency braking systems currently available, the company says.
Work is currently being carried out to integrate the distance to the vehicle behind the car into AHB’s activation strategy, with the aim being to reduce the risk of rear-end collisions caused by the car braking rapidly and the vehicle travelling behind it failing to take the appropriate braking or evasive action in time. The system can tailor the braking strategy in such a way that the car brakes earlier but less heavily, allowing the vehicle behind suitable warning of the car slowing.
Helping to avoid side-on collisions
For this one, there’s a new driver assistance system that goes by the name of Lateral Collision Avoidance (LCA), a further developement of Narrow Passage Assistant, a driver assistance system which helps the driver to follow the optimum central line along roads narrowed by roadworks.
The LCA is pretty much an extension of the Lane Change Warning System, which monitors the driver’s blind spot. The system, which features ultrasonic sensors positioned at the front and rear of the vehicle flanks to monitor the sides of the car, works on all roads with at least two lanes – the sensors monitor an area on either side of the car measuring anything up to four metres, depending on the car’s speed. In the current research prototype the system works up to speeds of 130 kph, though the speed band is surely to be extended.
If another vehicle enters a predefined area around the car, the driver is first alerted by a symbol in the HUD, and if the other vehicle closes to within a critical distance, the alert symbol becomes a warning, accompanied by a slight steering impulse that suggests to the driver to immediately move away from the danger. Tests show this form of feedback – which can be overidden – to the driver is immediately and intuitively understandable.
Look ma, no hands! Not
Sure, you’ve got the likes of Active Cruise Control with Stop & Go function (ACC Stop & Go), which not only maintains the desired distance to the vehicle in front but can brake the car down to a standstill in heavy traffic. But how about if the car could also actively control the steering, especially in more mundane situations such as traffic jams and queues?
That’s what the Traffic Jam and Queuing Assistant project is all about – a car that can think and steer, from standstill up to 130 kph. The system extends the reach of ACC Stop & Go to include a lateral guidance function – the difference is that the car can now also actively control the steering. Further development of camera tech enables the car to anticipate the path of the road ahead on the basis of the road markings and automatically carry out minor corrections in its line.
However, there are limits to how much steering work is possible – automated driving isn’t yet achievable – or desirable – through tight corners, so if a corner is too tight or the system reaches its limits due to insufficient road markings, the system prompts the driver to take over full driving duties again and switches off. The system, which complements the proximity control function of ACC by actively introducing steering inputs to correct the car’s line, is only active when the driver has his hands on the steering wheel. Hands off and the system automatically deactivates.
So why would one want something like this around? Well, up to certain speeds, this would allow the driver to work on emails or call up multimedia applications, though whether this is a good thing or not is another thing altogether.
And to help things, wear an AMULETT
Back on the pedestrian front, transponder-based systems such as AMULETT and Ko-TAG are helping to prevent contact between a vehicle and a pedestrian in the first place, especially in situations where pedestrians are concealed from view (coming out from behind a parked car, for example).
The AMULETT research project (a mouthful less than the full term, which is ‘Active mobile accident avoidance and mitigation of accident effects through cooperative data acquisition and tracking technology’) demonstrates “car-to-X” communication to improve pedestrian safety – a vehicle communicates with a wireless transponder worn or carried by the pedestrian or cyclist.
It works by transmitting a query signal by the vehicle, which the transponder then replies to with an identification signal. The system evaluates this reply, from which it is able to calculate the distance between the vehicle and the transponder, along with the angle of incidence of the signal, and it can also identify the type of road user. The EMV signal is evaluated by a 2.4 GHz multi-aerial system fitted behind the front windscreen of the test vehicle – based on the time lapse between transmission and reply, the system calculates the distance between vehicle and pedestrian on an echo time principle, offering a range of well over 100m in an unobstructed environment and around 20m if the pedestrian is concealed.
Yet again, the system works as mentioned earlier – detecting a potential collision situation, the driver gets a visual warning via the HUD as well as the central information display, and if he fails to react the warnings escalate right through to automatic emergency braking as a last-resort means of preventing a collision.
Working hand in hand with Ko-TAG
The findings of the AMULETT project are being taken further as part a research initiative called Ko-FAS – Cooperative Vehicle Safety, which is being carried out by BMW and 18 partners from throughout Germany. This will comprise development of a protocol to support system functionality in complex scenarios, along with further miniaturisation and industrialisation of the transponder.
Whereas at the start of the AMULETT project the transponder was still roughly the size of a school bag, the revised version is currently only the size of a small cigar box. In future, Ko-TAG transponders, as they are known, will shrink further still, and may soon even fit into a walking stick or inside a small compartment in a school bag.
A number of collateral issues exist. For example, the transponder cannot work without a reliable power supply. Also, reliable and interference-free data transmission requires a dedicated wireless frequency. Currently though, no frequencies are available, and it’s unlikely that any frequencies will be freed up for this type of safety function. Additionally, a transponder can only protect individuals who actually wear or carry it, so a stated project objective is to find ways of integrating the transponder to allow a wide as possible deployment, for example in mobile phones, school bags or running shoes.
Stopping safely when there’s no more driver control
Imagine the scenario and the ramifications: a car is travelling on the outside lane of a busy motorway when the driver suffers a heart attack and is no longer able to control his vehicle. Enter Emergency Stop Assistant to combat any untoward mishaps that will undoubtedly come up in such a case.
It’s an assistance system that activates an autonomous driving mode when it detects that the driver has a serious medical problem, and carries out a controlled emergency stop. In simple terms, the car switches on the hazard warning lights and manoeuvres carefully – taking into account the traffic around it – to the outer edge of the road, before drawing to a standstill. At the same time, an emergency call is sent out containing the data required to initiate the necessary medical and traffic-related assistance measures.
The ESA system is in place in a research prototype. In the test vehicle, a heart attack can be “simulated” at the touch of a button on the steering wheel. The technology then takes over full control of the car, keeping the car on the move within its lane and maintaining a safe distance to the vehicle in front, before eventually moving onto the hard shoulder – changing lane several times, if required – and braking to a halt.
Secure pinpointing of the vehicle within its lane and, above all, the reliable recognition of all vehicles and objects in its immediate area is achieved by bringing together various sensor technologies such as LIDAR, radar and camera recording on all sides of the vehicle as well as GPS, and the vehicle uses at least two different measuring principles in each direction to clearly establish its position.
The system pulls back from initiating a lane change if the manoeuvre will impede another driver. Only when it’s possible to change lanes without creating danger does the vehicle move into the new lane. On a test run in Munich, I got to sample the system’s workings, and it was plenty impressive, the car avoiding other vehicle mules used to depict moving traffic and working its way to a stop, safely and surely. Ford has something along the same lines, though that one uses a heart rate monitoring seat. Eventually, that system’s overall workings aren’t going to be very different than the one BMW has.
Some other choice cuts to end it
There were many other technologies on call at the Connected Drive showcase, of which I’ll just run through a couple very briefly in summing up this rather lengthy discourse.
One of them is called the Green Driving Assistant, which offers drivers a tool that also tells them their fuel consumption over each route and helps them to select one which will burn less fuel. It’s also able to calculate if the vehicle has sufficient fuel in the tank to reach the desired destination with the current driving style and route. A learning phase of around 500 km is all that is required to produce an adapted, vehicle- and driver-specific value which can be used by the GDA each time the driver is planning a new route.
The other is predictive navigation, in which artificial intelligence is used to teach the pathfinders how to learn. Somewhere in the future, navigation systems will be able to use these “learned” skills to predict the destination of a journey without the driver having to enter it beforehand, to give warnings of traffic jams and to reduce fuel consumption. A 3 Series converted by research engineers into a prototype has been able to predict – through its IleNa system – with a high level of probability (80%) where a journey is heading and which route has been chosen, without the driver inputting the information in advance.
To make the required predictions, the navigation system first has to get to know the driver and his regular routes. A secure profile is created for each driver in which past journey history is recorded. This will include not only destinations, short cuts and rat runs used en route, but also information such as the time of day and seat occupancy. Much of the data absorbed by the intelligent navigation system is useful not only for individual drivers, but for every user on the network. Simply put, information on traffic flow and fuel consumption learned can also be shared with other vehicles – the predictions for the road ahead become more precise, possible errors in the maps are corrected and forecasts for the traffic situation ahead are improved. Truly connected, it all is.
[zenphotopress number=999 album=2389]