Chandrayaan 2’s Moon illusion

The Vikram lander, which carried a small rover called Pragya, being prepared for launch earlier this year. ISRO lost contact with the lander during its descent to the lunar surface earlier this month. (credit: ISRO)

At the time this article is going to press no change has occurred in the status of Chandrayaan 2, India’s second mission to Moon. The last update was on September 10 when the Indian Space Research Organisation’s (ISRO) website stated that “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.” (Vikram is named after Vikram Sarabhai, the father on India’s space program.) India’s second mission to the Moon started on a cautious note and ended up with limited success. Chandrayaan 2’s journey for 48 days was challenging and demanding and ISRO successfully ensured that the craft would reach safely the Moon’s surface as per the plan.

The journey

Initially, ISRO was to launch the Chandrayaan 2 mission on July 15. For this purpose a 20-hour countdown had begun on July14. However, the launch was called off due a technical snag just one hour before the launch. Subsequently, an expert committee appointed by ISRO analyzed the cause for the snag and corrective action was taken. Finally, Chandrayaan 2 launch took place on July 22.

This mission has two major parts: the orbiter and a lander-rover system. The orbiter is a 2,379-kilogram satellite while lander weighs 1,471 kilogram and the rover weighs 27 kilograms. The designed life for the orbiter is one year. However, ISRO saved significant fuel during its journey to Moon and hence now expects that the orbiter may function for a period of seven years. It would be studying the lunar surface from an approximate distance of 100 kilometers. The lander-rover system was to be positioned on the Moon’s surface and was to operate for 14 Earth days, from sunrise to sunset on the Moon.

The mission was launched with orbiter and lander-rover as one cohesive system. On September 2, the lander, with the rover inside it, separated from the orbiter for its journey towards the surface of the Moon. A soft landing was attempted on September 7, however the communication was lost at a height of 2.1 kilometers above the Moon’s surface and, as per the orbiter imagery, it is safe to conclude that a hard landing of the lander did happen. Only ISRO can correctly identify the reasons for this failure.

Some 50 years ago, Apollo 11 had reached lunar orbit in just 51 hours and 49 minutes time, using a very powerful Saturn V rocket. However, the Indian mission took 48 days, mainly because India does not have a powerful rocket.

Chandrayaan 2, from launch through August 6, underwent five orbit-raising maneuvers which increased its apogee to around 150,000 kilometers. Subsequently, on August 14, Trans Lunar Insertion (TLI) look place, followed by the Lunar Orbit Insertion (LOI) maneuver on August 20. Once in orbit around the Moon, five maneuvers lowered its orbit. After the separation of the lander took place on September 2, the lander successfully performed deorbiting maneuvers on September 3 and 4, and the final landing was attempted on September 7. Some available details about this travel (in open source) are presented in form of a table below:

Date Orbit-raising manoeuvre Proposed Actual Deviation
24 July First 230 x 45162 km 230 X 45,163 km 1 km
26 July Second 250 x 54,689 km 251 X 54,829 km 1 x 140 km
29 July Third 268 x 71,558 km 276 x 71,792 km 8 x 234 km
02 Aug Fourth 248 x 90,229 km 276 x 71,792 km 29 x 757 km
06 Aug Fifth (final) 221 x 143,585 km 276 x 142,975 km 55 x 610 km
14 Aug Trans Lunar Insertion (TLI)
20 Aug Moon Entry 118 x 18,078 km 114 x 18,072 km 4 x 6 km
21 Aug Second Lunar orbit manoeuvre 121x 4303 km 118 x 4412 km 3×109 km
28 Aug Third 178 x 1,411 km 179 x 1412 km 1 x 1 km
30 Aug Fourth 126×164 km 124 x 164 km 2 km
01 Sep Fifth 114x 128 km 119 x 127 km -5 x 1 km
02 Sep Lander separated from Orbiter
02 Sep 119 x 127 km
De-orbiting manoeuvre
03 Sep first 109 x 120 km 104 x 128 km 5 x-8 km
04 Sep second 36 x 110 km 35 x 101 km -1x 9 km
07 Sep Hard Landing Occurred

This table indicates that on a few occasions there were significant deviations between proposed and actual orbits. But they could be considered less significant from the point of view of the health of the mission, if you consider the amount of the distance the orbiter was travelling.

Is it a failure of autonomy?

For any planetary exploration, landing on the surface of a planet is the most challenging part for the mission. The chairman of ISRO had actually pronounced on various occasions that the last part of the mission which involves the controlled (powered) landing is going to be terrorizing! This involved the travel of lander for last 35 kilometers. This travel was to take 15 minutes, and became known as the “15 minutes of terror.”

Everything was supposed to happen in autonomous mode during this 15-minute period of descent. The ISRO Chairman said, “We cannot do anything… We cannot interrupt. It is total autonomy. We are doing this for the first time.” A deviation in the trajectory, 12 minutes into the descent, was visible. Obviously, the system malfunctioned during the autonomous mode. Possibly, the velocity was higher than required at that height. There is also a view that the problem could have involved some kind of glitch in the functioning of engines used in the descent. All in all, whatever may be the problem, the system did not go in safe-mode anticipating the problems. Did this happen because the lander was very close to the landing site for the system to take any corrective measures?

It is not the purpose here to undertake any detailed failure assessment. Broadly, it could be inferred that the system was not trained adequately to identify the fault and take corrective measures. The strength of autonomy of any system would depend on the basic design factors. The design would be mainly based on what the scientists and technologists anticipate as possible anomalies and feed them in the system. Normally, autonomous systems operate in complex and open-ended environments with high levels of independence. They are expected to learn and reason with themselves. In addition, they are designed in such a way that they would identify the unforeseen changes in time and respond accordingly. What isn’t publicly known is the nature and quality of artificial intelligence (AI) involved in entire process of landing of the lander Vikram. The rover was supposed to be a totally AI-based product. However, since it did not have an opportunity to operate, no judgement could be made in that regard. It could be very difficult to quantify the role, if any, played by AI during last few minutes of the mission. AI works on data inputs and for such unique missions it is unlikely that required data would have been available in abundance. As such, globally certain limitations of AI are been constantly debated. It appears that space, which is a data sparse region, may not be a suitable region for keeping blind faith in AI.

The way ahead

Broadly, it could be said that ISRO led a very spirited Moon campaign, but failed to manage a soft-landing for its lander Vikram. In near future, after undertaking the detailed fault assessment, ISRO can think of repeating the mission (say, Chandrayaan 2R) by undertaking required corrections. Already they have an operational orbiter around the Moon. They know how handle deep space communications. Normally, space agencies develop two models of one equipment (one main and one spare). There is a possibility that ISRO could also have an additional pair of lander-rover systems available to try again.

Note: we are temporarily moderating all comments subcommitted to deal with a surge in spam.