Solar Sail Interplanetary Transfer: Part I - Earth-Mars Rendezvous Using Evolutionary Neural Network

Deep space exploration has long been a game of chemical rockets, massive engines that burn through fuel to push spacecraft forward, only to coast for the rest of the journey. But a different kind of propulsion has emerged, one that requires no fuel at all. Instead of burning propellant, these spacecraft use the pressure of sunlight itself. Imagine a giant, shimmering mirror floating in space; as photons from the sun strike its surface, they transfer a tiny amount of momentum, pushing the craft forward. This is the solar sail. While the concept has been tested in recent years by missions from Japan, the United States, and other organizations, turning this gentle, continuous push into a precise path to another planet remains a formidable challenge. The mathematics involved are incredibly complex, filled with non-linear twists that make finding the perfect route difficult for traditional computers. If a spacecraft misses its mark by even a small margin, it could drift past its target forever.

To solve this, researchers Renuganth Varatharajoo, Amirul Fiqri Abdullah, and Yury Razoumny have developed a new way to plan these journeys. They combined two powerful tools: an evolutionary algorithm, which mimics the process of natural selection to test thousands of possible solutions, and a neural network, a type of computer program inspired by the human brain that can learn to make decisions. In their study, they set out to see if this hybrid system could guide a solar sail from Earth to Mars. They did not build a physical ship; instead, they created a detailed digital simulation of the solar system. In this virtual environment, they tasked the computer with finding the best steering angles for the sail to reach Mars, arriving with the correct speed and position to meet the planet. The goal was to minimize the time it took to get there while ensuring the spacecraft did not fly too far past or too far behind its target.

The researchers tested their method under three different conditions, representing solar sails with varying levels of performance. The first scenario used a highly efficient sail, one that could generate a strong push. In this case, the computer successfully plotted a course that brought the spacecraft to Mars in just under 400 days. The second test used a sail with half the performance of the first. This was more difficult, requiring the computer to work harder to find a path, and the journey took nearly 600 days. The final and most challenging test involved a very low-performance sail, one that produces only a tiny fraction of the thrust of the first. Here, the journey stretched to nearly 2,400 days. These results showed that the system could find a solution even when the spacecraft was very weak, but it required significantly more time to reach the destination.

A key part of their discovery involved how the computer learned to navigate. The researchers found that the number of data points the computer used to check its progress mattered greatly. When they used a small number of checkpoints, the computer could find a path quickly, but the arrival was less precise. When they increased the number of checkpoints, the computer could make finer adjustments, improving the accuracy of the arrival speed. However, they discovered a trade-off. Increasing the number of checkpoints did not always make the overall mission better. In some cases, improving the speed at arrival meant the spacecraft arrived at a slightly worse distance from the planet, or the total travel time increased. The system seemed to prioritize one goal over another depending on how it was set up. This suggests that there is no single "perfect" setting for every mission; instead, mission planners must balance the need for precision against the available computing power and the specific requirements of the trip.

The study demonstrates that this evolutionary neural network approach is a viable way to solve the complex puzzle of solar sail navigation. By allowing the computer to explore a vast range of possibilities and learn from its mistakes, it can find routes that traditional methods might miss. The researchers showed that this method works well for a journey to Mars, even with a low-performance sail, and that it can adapt to different mission constraints. While the results are currently limited to computer simulations, the framework is designed to be flexible. It could potentially be adapted for missions to other planets or asteroids, offering a new tool for future explorers who wish to travel the solar system using the power of the sun. The work confirms that with the right computational tools, the gentle push of sunlight can be harnessed to guide a spacecraft across the vast distances between worlds.