@Article{2018.100000027,
AUTHOR = {Kaijun Xu, Xueting Chen, Yusheng Yao, Shanshan Li},
TITLE = {The Lateral Conflict Risk Assessment for Low-altitude Training Airspace Using Weakly Supervised Learning Method},
JOURNAL = {Intelligent Automation \& Soft Computing},
VOLUME = {24},
YEAR = {2018},
NUMBER = {3},
PAGES = {603--611},
URL = {http://www.techscience.com/iasc/v24n3/39785},
ISSN = {2326-005X},
ABSTRACT = {The lateral conflict risk assessment of low-altitude training airspace strategic 
planning, which is based on the TSE errors has always been a difficult task for 
training flight research. In order to effectively evaluate the safety interval and 
lateral collision risk in training airspace, in this paper, TSE error performance
using a weakly supervised learning method was modelled. First, the lateral 
probability density function of TSE is given by using a multidimensional random 
variable covariance matrix, and the risk model of a training flight lateral collision 
based on TSE error is established. The lateral conflict risk in specific training 
airspace is analyzed, and then the lateral collision model is built. Through the 
quantification of the risk probability of lateral collisions, the security level of a 
specific airspace is evaluated. The analysis of the examples shows that for normal 
training flight in a variety of 4D flight track data, the lateral collision risk in 
specific training airspace is  0.543744 × 10<sup>-13</sup>, the conflict risk meets the 
requirement of safety target level of international civil aviation organization.},
DOI = {10.31209/2018.100000027}
}