Where can this solution help?

• Hot spring and sauna facilities
• Nursing Homes
• Pool facilities
• school

R60ATR2

Infineon products installed
- 60GHz millimeter wave radar: BGT60TR13C

Features
- Detection distance: Max. 5m
・FOV (viewing angle): ±50°
Power supply voltage: 5V
- Current consumption: 93mA (estimated) *Customization to lower current consumption is possible
Size: 35 x 31 x 7.5mm

Infineon millimeter wave radar product page (Infineon website)

Evaluation board block diagram

Get the evaluation board

1. IF signal acquisition

First, the transmitted radio waves are reflected by various objects in the environment, and an IF (intermediate frequency) signal is obtained from the difference in frequency with the received signal. This signal includes not only reflections from the human body, but also reflections from walls, floors, and even water droplets and steam. Therefore, the signal at this point is heavily influenced by water, making it difficult to accurately detect human movement as it is.

2. 1D FFT (range FFT) for each chirp

Next, a one-dimensional FFT (Fast Fourier Transform) is applied to the acquired IF signal. This processing makes it possible to break down the signal into different frequency components for each distance. In other words, it becomes possible to identify the "distance from which each reflected wave returned." At this point, the reflected components affected by water and those affected by people are mixed together, so it is necessary to properly separate these in later processing.

3. Divide the FFT result by the number of samples and smooth it.

If the FFT results are used as is, environmental noise and small fluctuations will be emphasized, and the effects of water droplets and steam will be magnified. Therefore, by dividing the FFT results by the number of samples, the data is smoothed and extreme fluctuations are suppressed. This process results in smoother data, enabling stable analysis.

4. Velocity information analysis using Doppler FFT

After obtaining information for each distance using Range FFT, the next step is Doppler FFT (velocity analysis). By applying Doppler FFT, it is possible to measure whether an object is moving, in which direction, and at what speed. This processing makes it possible to more accurately distinguish between stationary objects such as water droplets and walls and the movement of people.

Specifically, the system analyzes the phase change over time over a certain distance, and determines whether the object is "stationary" or "moving" based on the magnitude of the change. This information is then used in later algorithms to detect falls and long periods of stationary motion.

5. Algorithm Processing

Based on the distance and speed information obtained through this processing, algorithm processing is performed to accurately detect a person's condition while minimizing the effects of water as much as possible.

Specifically, in environments such as bathrooms, shower water droplets and steam can easily affect radar signals, so filtering is done to take these effects into account and outliers are removed by comparing with past data. Furthermore, by detecting abnormal movements compared to normal human movements, it is possible to adjust the system to identify falls in the bathroom or long periods of immobility (loss of consciousness or other abnormal situations).

6. MTI Processing

Finally, MTI (Moving Target Indication) processing is applied, and filtering is performed to remove unwanted noise caused by stationary objects and water droplets. The main purpose of this processing is to emphasize "moving objects" and eliminate "stationary objects."

For example, bathroom walls, floors, and water droplets generally do not move, so these components are reduced and signals related mainly to "human movement" are extracted. This enables more accurate detection of people while minimizing the effects of water.