Technical Solutions
Given some sensitivity we are only sharing some high level view of our technical systems and overall architecture here. Some of the key sensors are missing from these diagrams and some flows we are keeping internally, but these show enough of our process that is starting to shape given all the traction, engineering and development shown above.
We have spent time planning this architecture as it has to be very precise and synced down to millisecond levels across the data objects. This in itself is a very big challenge then to sync the timelines across other activities with different calendar timelines.
The diagrams below show some of the sensors and the data fusion we are employing in our vehicles and simulators.
Our human feedback loops are some of the most important architectural aspects of our data pipeline. This image is a simplification of the architecture, but depicts our two main ML pipelines, our internal and our product based pipeline.
Engineering Shop & Labs
Our engineering shop is our main R&D facility that has taken us a year and a half to put together with all the tools and equipment capable of re-engineering and customizing our cars. Outfitting them with sensors and hardware without gaining considerable weight or altering the physics of the vehicle is important for the derived data.
Full Engineering shop
We got complete engineering shop setup with car lifts and all the tools complete to engineer our own vehicles and hardware. Plenty of floor space to expand and co-located labs and private datacenter.
Electronics Lab
We are setup to build and alter sensors and hardware in our lab, that is already stocked with parts and essentials.
Testing and QA Lab
We are setup to test sensors in simulations and NVIDIA Jetson systems. Electronic testing equipment and much more.
The high variance in quality vs pricing on hardware for sensors is a challenge on a tight budget but also important for outcome of the data and products, we rigorously test hardware and equipment from many vendors to find the right sensors for our use case. We also CAD and design some of our own hardware and sensor but use third party manufacturers and vendors to produce them.
Private Datacenter
Our private datacenter is taking shape with enough power supply to power our own small scale GPU and data storage. This is for R&D and the building of our models in-house and staging of our data before it is transferred to a major cloud provider, where larger tasks and compute is taking place.
Datacenter
Our own private datacenter for R&D and ultimate data privacy for our data is important to allow us to rapid prototype and bring products to market without relying on other vendors and supply chains and wait times.
Fiber Optical uplink
We operate 2 /24 CIDR IPv4 networks with our own networking and security systems using state of the art vendors, where are also experimenting with new modern cybersecurity solution vendors.
Power & Cooling
Underground power upgrade has been added to our building and full AC-DC-AC backup power systems and cooling systems is installed. We need to add generator for critical loads post funding.
Custom Car Engineering Progress
Our first car build is close to completion, this will serve as our first sensor deployment vehicle that is the beginning of pushing data into our pipeline.
Our second vehicle for data collection will be Subaru WRX STI modified with forged internals and stage 1 setup. Currently undergoing engine rebuilt and being outfitted with sensors. This will be a light sensor version that is targeting urban environment and roads-less-travelled, our target geographical areas are where there is no autonomous and the least EV presence to date.
CTRL-DRIVE are actively building 4 different cars some in later stage then others, each targeted for different type of data collection where our engineering efforts is based on purely edge-case data generation. This is a complex and counter intuitive engineering efforts where the purpose is not to engineer the car to loose traction but for the operator to push limits of what the mechanical limits are.
Simulator Labs
High-end simulators with VR and haptic feedbacks for G-Force emulations. A lot of engineering and code development has gone into the SIM setups to enable extraction of data, and modification of the simulator softwares itself. We are making extensive use of Assetto Corsa with our own mods and custom sensors and telemetry added.
Home Lab & Mobile Lab
Our home lab and development environment complements our primary engineering facility, enabling 24/7 R&D iteration with dedicated simulator rooms and a secondary compute cluster linked via 1Gbps to the main datacenter
Home Lab
A lot of hardware for testing and inference and simulation trials for R&D.
Home Datacenter
Our non motion simulators and software simulations connects to home datacenter for data staging and R&D in our home lab.
Mobile Lab (In the works)
We have assembled equipment but need seed funding for the trailer purchase.
Our mobile lab equipment is starting to be collected but the actual mobile lab will be purchased post seed funding. Our home lab is state of the art setup with proper security and networking, our home lab has dedicated power supply and proper cooling.
Simulations and Software testing
Simulations
We make use of NVIDIA AlpaSIM and other vendors simulations software to validate and develop our improvements towards simulations closing the sim-to-real gap with our products internally first.
Software Stacks
Our stack ranges from vehicles systems to mobile labs and datacenter internally using hundreds of open source solutions and commercial solutions, in addition to our own developed platforms for internal use.
ML / AI
We maintain many ML/AI tools for deep-research, engineering and our internal builds of specialized ML tools that is building our products in sync with humans and become teacher models for our customer facing product models.
We are using AI generated images here for now, until we are ready to share some actual video files collages from our Simulations, this takes efforts and will be handled once we have the bandwidth to scrub and sanitize them. For now they will available post NDA in Virtual Data Room.
Patent And IP
Our first 3 process patents are in the works and will be filed for patent pending Q2 2026, we anticipate more patents post funding from R&D and access to resources for research and writing.
Mastering All Aspects of Motorsports
Racing-to-Road is a loaded terminology when it comes to CTRL-DRIVE our view on this involves all possible tracks and physics we can encounter for our vehicles and our operators, our activities are tied to our innovation and depended on engineering and rapid prototyping.
Honda Racing Corporation (HRC) sharing track day with us at ButtonWillow in their prototype race car.
We are in conversations with partners to initiate our Rally data collection in Q3 2026 in Texas. We envision this is an activity that will yield high amounts of data insights across much shorter time of activity, both from physics and the human teaming portion.
Operators Teaching Moments
Operators teaching moments with human-to-human knowledge transfers. These are key data objects that plays a role in the overall data insights one some elements.
CTRL-DRIVE is pursuing partnerships and relationships across the motorsports industry to where we can leverage our setups and engineering to augment other teams and shops, in return for data generation. These relationships are already in discussions across varied motorsports verticals.
Beginning Of Aviation
Our aviation efforts are slated post Series A funding as it will require CAPEX and OPEX to operate a single plane that is capable of acrobatics and sports flying. But we are investing in our operators so they can operate on simulators and understand physics and theory behind simulations in the vertical.
Aviation data capture is a Series A milestone, not a seed priority. Current investment is limited to operator familiarization on simulators.
