The Star Bulletin - Daily News Updates

A new AI technology measures the noise of upstairs neighbors

A new AI technology measures the noise of upstairs neighbors

A new AI technology measures the noise of upstairs neighbors.

People can’t sleep well due to the noise of their upstairs neighbors. These sleepless nights are common in South Korea due to the loud upstairs neighbors. Living in an apartment unit means that you have to deal with the noise of the neighbors on a daily basis.

Korea Institute of Civil Engineering and Building Technology (KICT) President Kim ByungSuk has announced a new method for predicting the footsteps of upstairs residents. It uses a convolutional neural networks(CNN) model based upon vibration signals. CNN models are used in many computer vision tasks. To monitor the footstep-induced vibration, vibration sensors can be mounted on the walls and floors of residential buildings.

Floor noise from upstairs can cause stress for occupants and lead to conflict between neighbors. Korea Environment Corporation’s 2022 survey found that footsteps were the leading source of noise complaints in apartments. 

Scientists record the first sound recording of Mars dust devils

This accounts for 67.2%. Hammering was rated at 10.6% and furniture dragging at 5.5%. Heavy-weight impact sound is the main source of noise from neighbors in apartment units. These are mostly reinforced concrete box-frame structures. There is no way to get objective sound information.

A new AI technology measures the noise of upstairs neighbors

For heavy-weight impacts, numerical methods such as finite element analysis and statistical energy analysis can be used. It is not possible to predict sound when structure and constraint conditions are complex. Additionally, calculations of physics-based models can take several hours.

This algorithm predicts the impact sound of footsteps, particularly in rooms. Experimentally, a dataset was collected. It was then compared according to the location of vibration sensors and the resolution for the short-time Fourier transform feature (STFT), which is footstep-induced vibrations. With 0.99 dB being the mean absolute error, the sound level was predicted for 2 s.

There is not enough evidence to prove the sound level or who made it. Third parties that mediate inter-floor noise disturbances among neighbors must consider the subjective opinions of both the complainant and those who are suspected of making the noise.

A future ‘impact monitoring system’ which predicts sound using vibrations could be useful to change the neighbor’s behavior. It is also possible to use the stored data by mediators in the event of disputes to identify and assess the source household for the excessive sound.

Shin, the main researcher, noted that it was more important to record occupants’ perceptions of noise levels in their living spaces and inter-floor noise. This data can be used to identify neighbors who perceive noise as being too loud. Shin stated that it was important to determine the noise exposure of occupants in order to reduce noise problems between floors. AI-based technology will allow for effective monitoring of inter-floor noise, so that people won’t be as affected by noise from their neighbors in the future.

Korea Institute of Civil Engineering and Building Technology is a government-funded research institute that aims to develop practical technology and source technology in the areas of construction and land management.

By Patsy S. Nielsen

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts

No widgets found. Go to Widget page and add the widget in Offcanvas Sidebar Widget Area.