Child Presence Detection
Saving Children's Lives with Artificial Intelligence

Summer brings swimming pool weather and lots of sunshine, which is cause for joy for many people, but it also harbors dangers. Every year, young children around the world die from heat-related causes because they are left unattended in parked vehicles. In the USA alone, where these cases have been recorded systematically for years, well over 900 young children have died this way since 1998. Most of them, nearly 85 percent, are four years old or younger. Tragic deaths of this kind also occur regularly in Europe and other parts of the world.

Photo from the driver's seat to the back seat: a little girl sits with her teddy bear strapped into the child seat. — With strong sunlight, even low outside temperatures of 14 degrees are enough for critical situations to arise for small children inside the vehicle.

No heat records are necessary; even low outside temperatures of 14 degrees Celsius are enough to create critical situations for small children inside the vehicle. It is not without reason that the European New Car Assessment Programme (EU-RO NCAP) is planning binding specifications for improved interior protection for children in new vehicles.

UWB radar technology and artificial intelligence
measure and detect infant vital signs

The global mechatronics expert, Marquardt, is already one step ahead in this field. Led by communications tech-nology and innovation expert, Andreas Becher, from the company's own research and development department, a team has developed a system that reliably detects whether a child is in the vehicle. The system can recognize the vital signs of the occupants and accurately relay this information to the vehicle's electronics. By analyzing this data, the system can assess the occupant's state of health and, with the integration of additional sensor technolo-gy, even approximate their age. This breakthrough enables the triggering of emergency warnings when necessary, which can occur in several steps, depending on the vital information, and may ultimately involve forwarding the data to emergency services. In cases of emergencies, the system known as "Child Presence Detection" (CPD) automatically sends out warning signals at different escalation levels, up to alerting emergency services.

Marquardt project manager Andreas Becher (right) together with innovation manager Markus Kramer during a test of the Child Presence Detection System. Becher stands at a laptop, while Kramer stands in front of a screen with the measurement data.

"We have been working intensively on a solution since 2021. The result is a functional prototype that has proven that our system works reliably and already meets the expected EURO NCAP requirements," explains Markus Kramer, Head of Innovation at Marquardt. "The next step is to enter series development together with the automotive industry."

Artificial Intelligence analyses movement patterns

In its CPD system, the mechatronics specialist uses high-performance radar sensor technology that operates on the basis of ultra-wideband (UWB) radiation. This technology is harmless to the health of the occupants. For the evaluation of the radar data, the Marquardt developers relied on software algorithms with artificial intelligence (AI). The UWB radar, whose sensors are located in each row of seats in a vehicle, determines whether there are moving objects in the passenger compartment by measuring the time of flight of the reflected signal. The AI, in the form of self-learning algorithms, is able to distinguish and assign the movement patterns detected by the radar.

Marquardt innovation manager Markus Kramer sits in the cab of a passenger car with a little girl; he shows her a measuring device that Marquardt is using to further develop its AI-based technology to protect children's lives. — In a medically supervised series of tests with children between the ages of zero and ten, Marquardt obtained measurement data with which artificial intelligence was trained to protect children's lives.

Our solution is an innovation which can save lifes.

Markus Kramer, Head of Innovation

"In the precisely coordinated interaction of both technologies, our solution recognizes whether the detected movement originates from an object or a living being. With the help of several sensors, the determination of the age group can be carried out," emphasizes Andreas Becher. "This capability is important because an alarm only needs to be triggered if it detects that a baby or small child has been left alone in the vehicle. Especially in this age group, it can be assumed that they will not be able to extricate themselves from the emergency situation."

Human respiration as a measurand

To detect movements in the vehicle interior, the CPD system from Marquardt is based on human breathing. Unlike the movement of limbs, the raising and lowering of the rib cage is repeated permanently, even during sleep or fainting. Moreover, this happens very evenly and exhibits a specific pattern. In addition, breathing frequency differs depending on age. "Our system enables the vehicle with further sensor technology to assign the detected breathing movement to a specific age group andthus decide whether an alarm is triggered or not," explains Andreas Becher. Breathing movement can be reliably detected even through thick fabric or foreign objects such as installed child seats with approved blinds.

A little boy takes a photo from the driver's seat of a passenger car. He is sitting in his child seat buckled in the back seat. — Marquardt uses radar sensors and artificial intelligence to improve the protection of children in cars.

Specialized Medical Tests With Children

Marquardt obtained the relevant data in a medically supervised series of tests with over one hundred children between the ages of zero and ten. The breathing movements of the young test subjects were recorded in a test vehicle. A pediatrician with an additional technical degree worked with the development team to evaluate the measurement data and support the correct execution of the test. The findings obtained in this way form the basis for training the artificial intelligence so that it can correctly interpret the signal pattern sequence recorded by the UWB radar. "This holistic development approach, which encompasses both the relevant medical research and our technological expertise in radar sensors, UWB radio communication and algorithms acquired over many years, is what makes our solution so reliable, an innovation that can save lives," says Markus Kramer.