rFpro, originally rFactor Pro, is a driving simulation software used by racing teams and car manufacturers for advanced driver-assistance systems, self-driving cars and vehicle dynamics. rFactor Pro was created in 2007 as a project of a F1 racing team, using Image Space Incorporated's rFactor as a codebase.[1] It has since been used by more F1 racing teams, top road car OEMs, Tier 1 suppliers, and motorsport manufacturers.[2] It was originally developed for driver-in-the-Loop simulations,[3] but has since been used for autonomous vehicle training as well. It is not licensed to consumers.[4]
History
rFactor Pro was created in 2007 as a project of a F1 team, using the rFactor simulator as a codebase,[1] and has since been used by more F1 racing teams,[5][6] including Force India in 2009,[7]Ferrari in 2014[8][9] and Alfa Romeo in 2019.[10]
rFpro is developed by rFpro Limited, based in Wiltshire, UK.[2] In 2017 rFpro acquired Image Space Incorporated's ISIMotor gaming engine, including the gMotor graphics engine, which it had been licensing since 2007.[11] In 2019 rFpro was acquired by AB Dynamics.[10]
In 2020 rFpro partnered with cosin scientific software to enable FTire (Flexible Ring Tire Model) to run with rFpro.[12][13][14]
Features
rFpro features a 120 Hz graphics engine, a library of high definition laser scanned tracks and roads, and an infrastructure in which users can plug their in-house vehicle physics through a Simulink or a C/C++ interface.[15][1][16] Alternatively rFpro rigid multibody physics engine can be used, which samples suspension and drive-train at 800 Hz.[17]rFpro includes a tool called TerrainServer, which can feed the LiDAR data with a 1 cm resolution to a vehicle model running in realtime up to 5 kHz.[3] The library of laser scanned tracks includes most of those used in the F1 championship.[18]
In switching to rFpro for its simulator software in 2014, the Ferrari F1 team cited the high fidelity of the reproduced track surface, with an accuracy better than 1mm in Z (height) and 1 cm in X and Y (position), which represented a ten-fold improvement over their previous solution.[8] They also cited the ability to respond to dynamic inputs faster than the driver can detect.
