University of Birmingham
A systematic investigation of the applicability of datamining for the improvement of pedestrian safety
Abstract
dc:description.abstractExtensive research has shown advanced technology as a promising component towards improving the safety of pedestrians in an urban road environment. In GB, in 2017, evidence from a national database on road data (i.e., STATS 19) summed 1305 road fatalities, with pedestrian fatalities representing c.a. 25.7% of the total road fatalities. Datamining, the technique adopted in this study, is a process used to convert raw data into useful information. Fewer studies, however, have investigated a systematic approach adopting datamining techniques (e.g., ML) to assess its level of quality and consistency. Therefore, the applicability of datamining to reveal undiscovered pedestrian road-crash casualty patterns more efficiently was further explored. This work uses ML to examine casualty statistics during a one-year period in GB. The investigation includes data regarding pedestrian attributes, the road environment and its physical infrastructure. A key advantage of ML over conventional statistics is that with minimal human intervention, advance warnings are given independently when exposed to new data. The adopted technique here, explores discrete classification rules. From a range of classification algorithms and for the road data given, the FCM was applied for its proven degree of objectivity and response sensitivity. To this end, the results anticipate pedestrian road-crash casualties mainly for elderly male pedestrians during the day with fine weather, with permanent speed limits set at 30 mph, particularly when neither general nor pedestrian facilities are made available. New pedestrian safety measures may be recommended as in, appropriate publicity, introducing variable speed limits, and further research focusing on developing non- existing pedestrian safety models on a number of pedestrian physical facilities (e.g., footbridge, pelican).
Degree
thesis:*- Name dc:type.qualificationname
- d_ph
- Level dc:type.qualificationlevel
- d_ph
- Grantor dc:publisher.institution
- University of Birmingham
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Muchanga, Leila Cristina
- ORCID person:identifier.orcid
Subjects
dc:subject × 3Rights
- Language dc:language.dcterms_iso639_2
- eng