AI, computer logic will pilot Chandrayaan 3 lander onto moon on August 23, ISRO scientists can only observe

“Braking ended at around two km,” was the message radioed back to mission control from the Deep Space Network earth station located outside Bengaluru.

The visuals on the giant screen at the mission control room suggested that Vikram had gone out of control and crash landed on the surface of the moon. Still, the mission scientists hoped for some signal to emanate from the Vikram lander on the surface of the moon.

“Vikram Lander descent was as planned and normal performance was observed up to an altitude of 2.1 km. Subsequently, communication from Lander to the ground stations was lost. Data is being analyzed,” was the official communication issued by ISRO after it was clear that the lander mission on Chandrayaan 2 had ended abruptly on September 7.

In the following days there was still hope that some signal would emanate from Vikram despite the crash. Nearly five days after the attempted landing ISRO announced on September 12 that the Chandrayaan 2 orbiter that was traveling around the moon had spotted the Vikram lander on the moon by using its Optical High Resolution Camera.

“Vikram lander has been located by the orbiter of Chandrayaan-2, but no communication with it yet. All possible efforts are being made to establish communication with lander,” ISRO said on September 12, 2019 even as senior retired ISRO scientists familiar with the efforts to establish a link with Vikram said “there is nothing to track — no communication signals.”

While mission control at ISTRAC and the mission scientists – including mission director P Veeramuthuvel – will keep a close watch on the Chandrayaan 3 landing effort on the moon on August 23, 2023 they will be able to make no interventions if things go wrong since the entire landing process is controlled during the 15 minutes of descent by computer logic that is already fed into the lander’s computers, guidance and control navigation systems.

“We will be closely monitoring and controlling the spacecraft from ISTRAC Bengaluru,” a nervous Chandrayaan 3 project director P Veeramuthuvel stated on July 14 when the mission was launched from the Satish Dhawan Space Center at Sriharikota. He indicated that the “final powered descent phase for soft landing,” is the key to the mission. The mission control in Bengaluru will receive data on signals sent by the Chandrayaan 3 lander directly to ground stations at the Deep Space Network at Bengaluru, the Jet Propulsion Laboratory in the US and a European Space Agency station in Spain or via the Chandrayaan 2 orbiter with which Chandrayaan 3 established contact on Monday, according to ISRO.

Mission control cannot however send commands to the lander once the descent phase begins on the evening of August 23 at 17:47 hours for a landing at 18:04 hours. In essence the Chandrayaan 3 lander will have to use its programmed AI to make a “safe and soft landing”.

“The core of Chandrayaan 3 is its sensors. When you have something that is remotely operated then everything depends on its ability to sense its location, what it is its speed, what is the orientation. There are different sensors used for this purpose. There are the velocimeters and altimeters which give a reference for the speed and the height of the lander,” ISRO chairman S Somnath said recently.

“There is a camera called a hazard avoidance camera, there are also inertia based cameras. The sensors are fused together using a computer algorithm to provide an indication of where the lander is positioned. This has been tested extensively,” he said.

According to the ISRO chairman the other key AI system is the navigation, guidance and control system on board the lander. “This is nothing but computer logic residing in Chandrayaan 3. This is what is steering it to the right position for landing. Whenever the lander has to take decisions on how to go it is the guidance and control system that will take the lander on the path for landing,” Somnath said.

“The configuration is designed in such a manner that we will have a safe and soft landing. We have addressed various things to achieve this goal of a safe and soft landing. We could not do it last time and we have addressed the issues that we faced in great detail,” he said.

“Extensive simulations have been done, the guidance designs have been changed, a lot of the algorithms have been put together to ensure that the required dispersions are obtained in all these phases. Even if there are variations in the nominal numbers still the lander will make an attempt at a vertical landing,” the ISRO chairman said.

“If all sensors fail, if everything fails it will still make a landing provided the propulsion system works well. This is how it has been designed. Even if two of the engines do not work also this time the lander will be able to land. It has been designed in such a way that it should be able to handle multiple failures. If the algorithms work well we should be able to do a vertical landing,” he said.

The onboard sensors will carry out calculations as the lander descends from a height of 30 km above the surface of the moon to a height of 7.42 KM in the nearly 10 minute first phase of the 15 minute landing effort.

“At a height of 800 or 1300 metres it will start doing a verification of the sensors, at 150 metres it will do a hazard verification and decide whether it should land vertically there itself or move laterally to a maximum extent of 150 metres to avoid any boulders or craters. Such choices and decisions will be taken at the appropriate time (by the lander),” Somnath said.