A Study on Wearable Airbag System Applied with Convolutional Neural Networks for Safety of Motorcycle

Woo, Joo; Jo, So‐Hyeon; Jeong, Jae‐Hoon; Min, Kim; Byun, Gi‐Sig

doi:10.1007/s42835-020-00353-5

Cited by 6 publications

(16 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most significant difference between the accident situation and the non-accident situation is based on which one of the acceleration and angular velocity axes varies and how much it varies in a short time. Depending on the information derived from the posture type and the extent of the impact, the values indicated by the acceleration and angular velocity axes may differ, and a similar value or outcome appears in a similar situation, which can be also confirmed through the precedent studies [ 26 , 27 , 28 , 29 , 32 ]. Figure 3 a,b shows similar graphs in the same accident situation (left collision).…”

Section: Methods and Design Of The Proposed Systemmentioning

confidence: 84%

“…This study collected the accident data on the above items to be used as data on the accident situations and explains the analysis of the cases depending on the situations. However, because it is very dangerous for people to directly ride a bike and make an accident, a mannequin was used in place of people, as shown in Figure 1 [ 26 , 27 , 28 , 29 ].…”

Section: Methods and Design Of The Proposed Systemmentioning

confidence: 99%

“…Complementary and Kalman filters are usually used extensively, but the complementary filter was chosen because the processing of the Kalman filter took longer than that of the complementary filter. To prevent accumulated errors due to the integral calculation, the weights of acceleration and angle data were varied and corrected according to the stopping and driving conditions [ 26 , 27 , 28 , 29 , 30 , 31 ]. When it is close to a stationary state, the weight of acceleration is increased, and when a sudden rotation or movement is detected, the weight of angular velocity is increased to respond to the changed angle.…”

Section: Methods and Design Of The Proposed Systemmentioning

confidence: 99%

“…Since there is a difference in characteristics of acceleration and angular velocity depending on each situation, the accident situation and non-accident situation were supposed to be determined through artificial intelligence learning by using this difference [ 32 ]. The experiments to collect data were conducted 20 times according to each situation, and the measurement period was 50 ms [ 26 , 27 , 28 , 29 ]. A total of about 658 data were collected, including non-accident data.…”

Section: Methods and Design Of The Proposed Systemmentioning

confidence: 99%

“…In the case of airbags, conventional products often inflate the airbag through a mechanical method or use the threshold value of a sensor, and 2–3 types of sensors are used together. Studies are underway to find the possibility of using artificial neural networks, instead of the threshold method used in the existing bike airbags, to produce an airbag system with high accuracy [ 26 , 27 , 28 , 29 ]. The precedent studies using other artificial intelligence models made judgments only with respect to the “incoming input data” and outputted the results.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Study on the Application of LSTM to Judge Bike Accidents for Inflating Wearable Airbags

Woo

Byun

et al. 2021

Sensors

Self Cite

View full text Add to dashboard Cite

The traffic accident occurrence rate is increasing relative to the increase in the number of people using personal mobility device (PM). This paper proposes an airbag system with a more efficient algorithm to decide the deployment of a wearable bike airbag in case of an accident. The existing wearable airbags are operated by judging the accident situations using the thresholds of sensors. However, in this case, the judgment accuracy can drop against various motions. This study used the long short-term memory (LSTM) model using the sensor values of the inertial measurement unit (IMU) as input values to judge accident occurrences, which obtains data in real time from the three acceleration-axis and three angular velocity-axis sensors on the driver motion states and judges whether or not an accident has occurred using the obtained data. The existing neural network (NN) or convolutional neural network (CNN) model judges only the input data. This study confirmed that this model has a higher judgment accuracy than the existing NN or CNN by giving strong points even in “past information” through LSTM by regarding the driver motion as time-series data.

show abstract

Section: Methods and Design Of The Proposed Systemmentioning

confidence: 84%