Overfilled roads and human driver errors rapidly cause traffic jams. A smart ZF system to improve partially automated driving could be an effective way to relieve strain and to enhance safety.
July 12, 2019
Kathrin Wildemann has been a part of the permanent Copy Team at ZF since 2016. In her online and offline articles, she likes to cover electromobility and other topics that involve sustainability.
The same picture across the board: tightly packed nose-to-tail traffic forming up every morning and evening on roads around the world – a slow-moving avalanche of sheet metal. All it takes to create gridlock are a few poorly judged overtaking maneuvers or a minor rear-end collision. Chronically congested transport arteries combined with human error are a sure recipe for traffic jams. To reduce the number of traffic jams, transportation experts placed great hopes in autonomous driving. A Herculean task. To find viable solutions to the collapse of road traffic that are suitable for daily use, diverse areas of application call for different levels of autonomous driving functionality.
New Mobility Concepts Keep the Traffic Moving
The most important way to combat congestion is to reduce the volume of traffic on the roads. In cities in particular, the roads and parking spaces have been overwhelmed for a long time, yet the number of vehicles continues to increase. A new mobility concept could provide a remedy to all of this: Mobility-as-a-Service (MaaS) for personal transport and Transport-as-a-Service providers (TaaS) for the transportation of goods. The backbone for this approach is formed by autonomous driving people movers and cargo movers. One example of this type of vehicle is the e.GO Mover that ZF has developed and produced in a joint venture with the company e.GO Mobile AG. The e.GO Mover is based on an electrically powered platform and is equipped with all the systems needed for Level 4 highly automated driving. It can be used as a passenger shuttle and as an LCV for the delivery of goods.
In future, autonomous people movers and cargo movers will be able to relieve strain on the road traffic system.
MaaS and TaaS vehicles do not follow any fixed schedule. Instead, they use an app to bundle incoming orders in real time. On this basis, they calculate the optimum route to get as many passengers or goods deliveries as possible to their destination using the shortest possible route. Ride-hailing services employ a comparable concept, using autonomous robo taxis. These smart vehicles reduce the volume of private transport. An example from the city of Munich illustrates this point. A study commissioned from Berylls Strategy Advisors, a strategy consultancy specializing in the automotive industry, indicates that a fleet of 18,000 autonomously driving taxis could replace about 200,000 private vehicles.
Statements like this encourage traffic planners, politicians, and business to drive forward autonomous driving as a problem-solver. Forecasts by Goldmann Sachs, Roland Berger, or McKinsey suggest that autonomously driving people movers and cargo movers could represent a global market potential of between 20 and 50 billion US dollars in 2030. Ride-hailing has an estimated future value of between 18 and 35 billion US dollars.
Autonomous ride-hailing vehicles do not follow a fixed schedule but instead use an app to bundle incoming orders in real time.
Costly Full Automation of the Passenger Car Is Not a Panacea
However, dispensing with drivers altogether does not constitute a universal solution either. From Level 3 onward – where drivers can divert their attention temporarily from the road ahead and carry out other tasks – the technology required becomes a great deal more complex and, by extension, much more expensive. As soon as total control is handed over to the vehicle, even for short periods of time, the requirements on the reliability of sensor systems become a great deal more demanding. Ultimately, the car must be able to identify every driving situation correctly and completely, irrespective of lighting conditions, weather, and speed. To satisfy this requirement, several cameras, radar sensors, and lidar sensors need to verify the measurement data independently of one another. This makes the system architecture much more elaborate and costly. Furthermore, rather than building on existing advanced driver assistance systems, it needs to have new algorithms and a massive increase in computing power. Purely on grounds of cost, this makes autonomous driving uninteresting to most private owners of passenger cars.
However, at the present time, these systems are frequently not synchronized with one another well enough, or shut down too rapidly in response to unfavorable ambient conditions such as heavy precipitation. As a result, they frequently cause more disappointment than relief to the driver. The key here then is to extend Level 2 functionality and to integrate it in a smart overall system that provides the driver with a uniform interface. With its ZF coPILOT system, ZF has presented a Level 2+ concept of this kind: the AI-capable mainframe computer ZF ProAI takes full charge over the control of all ADAS algorithms and networks them using a comprehensive set of sensors. This enables the vehicle to master driving and safety functions that extend well beyond the scope of Level 2 systems. For example, ZF coPILOT enables vehicles to merge autonomously with interstates, to change lanes, or to overtake. When a system of this kind takes the wheel in slow-moving evening commuter traffic, hectic overtaking maneuvers or unnecessary sharp braking cease to happen. The system is also able, using an intelligent navigation system, to identify an alternative route in plenty of time to avoid a location where congestion is anticipated. And what happens if the traffic suddenly comes to a standstill? At least drivers can then lean back in their seats relaxed and let their vehicles take care of that nerve-racking stop-and-go traffic.
The decisive factor is to extend Level 2 functionality and to integrate a smart overall system that provides the driver with a uniform interface.