RideFlux expands operations and increases ODD coverage by 5x

RideFlux develops self-driving technology to bring safe and efficient transportation into our daily lives. The company provides an end-to-end software solution for autonomous systems that is currently deployed in three commercial robotaxi services on Jeju Island, South Korea. One of these services is Korea’s longest route for commercial self-driving robotaxis, shuttling travelers between Jeju International Airport and Jungmun Tourist Complex for 38km each way. Another is Korea’s first free-floating autonomous mobility service, operating in Seogwipo Innovation City. Through its self-driving technology, RideFlux aims to reduce the need for individual car ownership and alleviate traffic congestion in South Korea.
Jeju-si, South Korea
(in 2022)
“Level 4 self-driving service technology requires driving on roads with high complexity and high uncertainty. Therefore, we need a dependable partner like Applied Intuition to quickly secure reliable autonomous driving technology.”
Junghee Park
Co-Founder and CEO
RideFlux seeks to scale its self-driving technology development, expand its operational design domain (ODD), and improve the stability and reliability of its robotaxi services to a level where safety drivers are no longer required (Level 4 autonomy). Before using Applied’s software, RideFlux developed its log data management and simulation tools in-house. This became challenging when RideFlux wanted to:
Manage log data effectively: The RideFlux fleet started with just one test vehicle, but the team quickly needed to operate, collect logs, and resolve issues on multiple robotaxis at a time. As RideFlux discovered limitations with its in-house tooling, it started looking for third-party tools that would be able to scale and automate workflows for its growing fleet.
Scale scenario creation: Along with its growing fleet of robotaxis, RideFlux decided to expand its ODD. Soon, the team’s in-house simulation tool could no longer create the large volume, variety, and complexity of scenarios needed to cover its ODD. Additionally, the scenario creation process was too manual and time-consuming to scale effectively.
Inside one of RideFlux’s self-driving robotaxis.
“Applied’s team provides high-quality software products and responds quickly to our requests.”
Junghee Park
Co-Founder and CEO
RideFlux decided to accelerate its development cycle and achieve its goals by focusing primarily on self-driving software development while leveraging third-party data management and simulation tools. The team chose to work with Applied Intuition because Applied’s tools facilitate efficient workflows for vehicle and simulation tests while being interoperable with Robot Operating System (ROS) and standard maps such as OpenDrive.

RideFlux uses Applied’s data management and simulation platforms Strada, Simian, and Logstream to:
Manage log data: RideFlux uses Strada to efficiently manage drive logs from its growing fleet of robotaxis. Strada also allows the team to conveniently filter and search its entire drive log collection for specific events of interest.
Root-cause issues: Strada’s user interface (UI) makes it easy to understand log data, identify the cause of an issue, and resolve issues without needing to open large raw data files.
Run re-simulation tests based on real-world data: Logstream enables RideFlux to create test cases from drive data. Using re-simulation, the engineering team can debug on-road events with the autonomy stack in the loop, test different versions of the stack on past drives, and catch regressions before they get deployed to the vehicle.
Create synthetic simulation scenarios: Simian’s easy-to-use UI allows the RideFlux team to create scenarios faster than its in-house simulator. It also allows the team to specify intelligent actor behaviors and automatically generate multiple scenario variations from one base scenario to build ODD coverage on scenarios that weren’t captured by drive data or re-simulation.

Automate tests: Through Simian, RideFlux can easily define and customize test criteria and set automatic checks to quickly determine whether a test passed or failed.
Validate software updates: With each new software update, RideFlux tracks changes, catches regressions on existing scenarios, and compares test results with the performance of previous software versions.
Viewing drive data from an intersection using Strada.
More efficient log data management
Using Applied’s software, RideFlux can now efficiently collect, manage, and resolve issues on its growing fleet of robotaxis. The team can dedicate a smaller headcount to these workflows than when it was building and maintaining tools in-house.
Faster development cycle
Applied’s software empowers RideFlux developers to quickly evaluate simulation and vehicle tests, easily share results with their team members, and leverage re-simulation to validate fixes. This improved workflow has helped the RideFlux team identify and solve a larger amount of issues, decrease its issue resolution timeline, and thus speed up its development cycle.
Increased ODD coverage
Since using Simian and Logstream, RideFlux has grown its scenario library by over 3x and increased coverage by over 5x while expanding its ODD.