Real-time trip information service for a large taxi fleet

Abstract

In this paper, we describe the design, analysis, implementation, and operational deployment of a real-time trip information system that provides passengers with the expected fare and trip duration of the taxi ride they are planning to take. This system was built in cooperation with a taxi operator that operates more than 15,000 taxis in Singapore. We first describe the overall system design and then explain the efficient algorithms used to achieve our predictions based on up to 21 months of historical data consisting of approximately 250 million paid taxi trips. We then describe various optimisations (involving region sizes, amount of history, and data mining techniques) and accuracy analysis (involving routes and weather) we performed to increase both the runtime performance and prediction accuracy. Our large scale evaluation demonstrates that our system is (a) accurate -with the mean fare error under 1 Singapore dollar (≈ 0.76 US$) and the mean duration error under three minutes, and (b) capable of real-time performance, processing thousands to millions of queries per second. Finally, we describe the lessons learned during the process of deploying this system into a production environment.

Lowering the barriers to large-scale mobile crowdsensing

“…In today's mobile device market, there are at least three popular software platforms, includ- In a sense, the complexity of the crowdsensing application space grows with cross product of the number of platforms and the number of applications. This issue of heterogeneity is underscored by the experience of Balan et al [2], who conducted one of the largest crowdsensing studies to date. It took them six months to deploy one version of their GPS-based crowdsensing application on 15,000 taxis in Singapore, mainly due to the heterogeneity of the on-car GPS devices provided by different vendors.…”

“…from social networking applications such as Twitter) that larger crowds have been studied. The one notable exception is the work of Balan et al [2], discussed in Section 2.…”

CallCab: A unified recommendation system for carpooling and regular taxicab services

“…For example, taxicab GPS records are able to help taxicab operators better oversee taxicabs and provide timely services to passengers, e.g., discovering temporal and spatial causal interactions to provide timely and efficient services in certain areas with disequilibrium [6] [7], and detecting anomalous taxicab trips to discover driver fraud or road network changes [9]. In addition to taxicab operators, several systems are proposed for the benefit of passengers or drivers, e.g., allowing taxicab passengers to query the expected duration and fare of a planed trip based on previous trips [3] and query real-time taxicab availability to make informed transportation choices [4], and recommending optimal pickup locations or routes [5]. Moveover, taxicab GPS records can help beyond the taxicab business: (i) traces consisting of GP-S records from experienced taxicab drivers can assist other drivers improve their driving performance [10]; (ii) GPS records can be used for navigating newer drivers to smart routes based on those of experienced taxicab drivers [11]; (iii) large scale taxicab GPS traces enable us to better understand traffic conditions of cities [12].…”

“…Unfortunately, almost all taxicab recommendation systems [3] [4] [5] [6] [7] are focused on how to find a vacant taxicab. Little work, if any, is focused on how to find a carpoolable taxicab for a passenger.…”

Towards Data-Driven On-Demand Transport

“…Email: malcom.egan@insa-lyon.fr such as the time and location of each passenger's pickup and drop-off, as well as the prices that passengers are prepared to pay and the fares drivers are willing to accept. The combination of this data is leading to new insights into the spatial-temporal profile of on-demand transport systems [1][2][3][4] and ultimately opens the way for increasingly efficient allocation and pricing.…”