Abnormal Condition Detection System based on Sensing / Analysis of Snow Removal Operations

Snow removal operations by snowplows play an important role for securing social activities and transportation of local residents in snowy and cold regions in Japan. However, there are situations where the operation must be suspended due to the heavy traffic of cars and pedestrians even at night. In order to improve the efficiency of the snow removal operations, it is necessary to take measures by identifying the areas where such situations are likely to occur. Therefore, in our study, we propose a new system which detects occurrence of the condition where the snow removal operation changes to the abnormal state. The proposed system uses several cameras to observe the motion of the operator to steer the snowplow. In addition, we propose a method to identify the date, time, and location that the motions of the snow removal operations are much different from the usual ones by analyzing the time-series data of the motion to steer. In our study, the proposed system is deployed in the snowplow managed by Hakuba Village of Nagano Prefecture, Japan to observe the maneuvering behavior of the snow removal operator and we evaluate the effectiveness of the proposed method for estimating abnormal conditions during snow removal operations.