Jack Trigger Racing.
Digital Twin of a High-Performance Dynamic System.

Case Study

Modern ocean racing sailing boats are high performance machines, almost more comparable to aircraft than the yachts of old. They combine cutting edge material science, aero and hydrodynamics, navigation systems, telecommunications, and sensors.

However, one underdeveloped technological domain is the use of real-time data from boat sensors for automated performance optimisation. While data is relayed via displays to a human, there is little or no interfacing with the autopilots that are vital for long distance racing with only one or two crew.

T-DAB.AI helped Jack Trigger Racing develop a Digital Twin to improve their autopilot function.

Trigger Racing Logo


Jack Trigger Racing (JTR) wanted to improve their autopilot function in their high-performance racing yacht using the data they had collected. To improve the automation function in this data-driven way requires a lot of trial and error and using a real boat for the experiments is not economically viable (also not safe) so there needs to be an environment which would simulate the boat’s behaviour in different race settings and under different weather conditions.

However, the boat is a 6-degree-of-freedom system with multiple parameters and complex aerodynamic and hydrodynamic features. Traditional modelling can take months of development and require experts in several disciplines, so JTR needed a low cost, yet effective means of simulation.


To take advantage of the wealth of data available from the boat sensors, T-DAB’s Innovation Lab team worked with JTR to develop a data-driven digital twin. Based on long short-term memory (LSTM) recurrent neural networks, the Digital Twin would observe the history of the boat’s dynamics and current weather conditions to forecast all the relevant parameters of the boat in the next second.

The function of the Digital Twin would provide the boat’s velocity and orientation in the given circumstances, which allows us to assess scenarios the boat has never actually been in.

This information can then be used to understand how the boat would perform and devise better approaches to controlling autopilot functions. It would also allow for the autopilot system to be exposed to extreme and often dangerous circumstances without endangering the physical boat or anyone on it.


The Digital Twin now allows JTR to receive information about the system very quickly and assess the consequences of autopilot actions. The first prototype of the LSTM-based twin of the yacht was developed quickly to assess the validity of further autopilot development.

Now that it has proven it’s potential, the digital twin can be developed further to increase system precision. For example, Recurrent neural networks showed a reasonable performance (within 10% of real state values) and has further potential for improvement through additional architectures such as convolutional and generative nets.

The Digital Twin enables a more cost effective and rapid prototyping in the pursuit of industrial automation and optimization. In the case of JTR, the digital twin facilitated quick progress in optimal steering and adaptability to new conditions, with the potential to improve power consumption and extend hardware life.

The enhanced features enable the JTR team to edge ever closer to seizing the winning advantage at the most demanding race challenges, such as the Vendee Globe.

to Find out more about the project,
read the full case study and get in touch with the team.

innovation at the edge

Read more about the full project and explore how we deliver research for application to deliver machine learning at the edge.