Development of a Smart Transportation Application to Improve Urban Public Transportation Efficiency in Kendari City

Unscheduled paratransit networks in developing cities, such as Kendari, frequently suffer from limited route visibility, severe arrival time uncertainties, and the lack of real-time monitoring systems, heavily eroding their public competitiveness. This study addresses these inefficiencies by developing Pete2Mi, a mobile based intelligent transportation application featuring a decentralized, bidirectional Location Based Service (LBS) architecture. Adopting the Design Research Methodology (DRM) integrated with an Agile development framework, the system leverages Global Positioning System (GPS) streams optimized via a recursive Kalman Filter algorithm to suppress multi-path urban noise. Estimated Time of Arrival (ETA) is modeled using low latency kinematic formulations. Empirical results demonstrate that the Kalman Filter significantly stabilizes tracking, reducing localization errors to a mean MAE of 1.63 meters (a 79.19% reduction relative to raw GPS data). Furthermore, human centric evaluations yielded an exceptional System Usability Scale (SUS) score of 81 (Grade A, Excellent) and an 86.98% acceptance rate via the USE framework. These findings validate that the bidirectional tracking paradigm successfully bridges the operational gap between informal paratransit systems and modern digital mobility ecosystems.