Pragyan rover roams around Shiv Shakti Point in pursuit of lunar secrets at the South Pole.Screen grab from ISRO video released on August 26
Artificial Intelligence may have been the key differentiator between the first 2 Chandrayaan missions -- and the spectacular third

Artificial Intelligence  may have been  the key differentiator between the first 2 unsuccessful  Chandrayaan missions -- and the  spectacular third


By Anand Parthasarathy

August 31:  2023: What sets apart the successful August 23   soft landing on the moon of  Chandrayaan-3’s  Vikram lander   and the subsequent investigations of the  Pragyan  robotic rover from  Chandrayan  1 and 2  in 2009 and 2019? 
There is general agreement that what played a part in the precisely executed manoeuvres was the heightened use of Artificial Intelligence in key operations like guidance and control.
Yes, AI was around in 2019 too – but there is no denying that the technology has matured  spectacularly in  the last 3-4 years in areas that could be directly deployed into the Chandrayaan programme, like:
Computer vision where AI enhances computers as they analyze digital images and classify objects, individuals, and actions. Recent advancements  have empowered robots  like Pragyan to achieve near -human-level performance in tasks like object detection and collision avoidance. Such roving robotic vehicles can interpret   large volumes of  visual data  in real time.
AI-optimized hardware  means specialized hardware architectures or components built and optimized to accelerate AI and machine learning (ML) applications.  When AI is harnessed to enhance graphical processing  units (GPU)  or central processing units (CPU), it can substantially improve the  quality of the outcomes.
Once the  Vikram lander went into its descent phase during the final  quarter-hour on August 23,  its ability to make a soft landing depended on  navigation and control: a  combination of its instantaneous location, speed, orientation. The instruments that achieved this  were velocimeters judging the speed,  and altimeters monitoring the height. Computers – enhanced by AI--  made the vital micro corrections which  ensured a  smooth touch down. Cameras on both orbiter and lander  -- the Lander Position Detection Camera (LPDC) and the Lander Hazard Detection & Avoidance Camera (LHDAC --harnessed AI-enhanced computer vision to  subtly change the lander’s  orientation from horizontal to vertical as it  to found the perfect landing spot. ISRO chairman S. Somnath,  was quoted  saying this was one of the most complicated manoeuvres of the descent.
Pragyan deployed critical made-in-India instruments
Once the Pragyan Rover started its autonomous journey on the moon’s surface,  AI would have kicked in again to ensure  that it could  navigate safely,   along the optimal path,   mapping the lunar features it encountered, even as it logged a mass of  critical data. It deploys indigenously developed instruments,  including a Laser Induced Breakdown Spectroscope (LIBS)  and an Alpha Particle Induced X-ray Spectroscope (APIXS) to  assess the composition of the elements  on the lunar surface.
It goes without saying that AI was  harnessed in the pre-flight stages when   multiple simulations helped scientists iron out any kinks – and post flight when  terabytes of data  need to be analysed   and processed.
Some of the above insights  came during the post -landing briefings by  ISRO – and if they were fairly intuitive and somewhat  general, that is because no SpaceTech player  is about to share  anything more than a broad  acknowledgment of its proprietary processes and software.
But as the corny saying goes,  the proof  is in the pudding, and  India’s demonstration of its  abilities only underlines the robustness and  cutting-edge quality  of its  indigenous space technology  during  a time when  AI  is increasingly proving to be an agni astra.

The article  appeared in Swarajya yesterday