IIT Madras develops motion planning algorithm that can think like humans

According to the team, the algorithm has been developed on a new notion of ‘Generalized Size Expansion’ (GSE) that enables planning of a safe and dynamically feasible trajectory for autonomous vehicles.

These approaches have been found to yield better results than existing fundamental and state-of-the-art motion planning algorithms.

The team claimed that due to its new calculation of the “safe” zone, it is a significant advance during time-sensitive planning scenarios arising in applications such as self-driving cars, disaster response, ISR operations, aerial drone delivery and planetary exploration. provides. ,

Unmanned Aerial Vehicles (UAVs) are often deployed to survey affected areas and scan debris for search and rescue missions. Since in such applications, UAV paths need to be planned in advance in a time-critical manner, these algorithms could play an important role, he said.

Satadal Ghosh, Assistant Professor, Department of Aerospace Engineering, IIT Madras, has published several research papers in internationally reputed peer-reviewed journals such as the AIAA Journal of Guidance, Control, and Dynamics, and IEEE Control System Letters, and top-notch journals. have done. Level conferences such as the IEEE Conference on Decision and Control (CDC), American Control Conference (ACC) and AIAA ScienceTech.

The team included IIT Madras alumnus Vrishabha Zinage, a doctoral research scholar at the University of Texas Austin (USA), Adhavait Ramkumar, a graduate student at Warsaw University of Technology, Poland, and Goldman Sachs analyst Nikhil P.

“The GSE-based algorithm works by computing a ‘safe’ area consisting of large ‘visible’ areas in the environment, which are optimized for navigating,” Zinez said.

“Then, algorithms select a random point in this ‘visible’ region and connect it via a secure ‘edge’ to the securely reachable areas discovered so far. After all, algorithms almost always have a point in any environment.” can also connect two points that meet some basic criteria,” Zinez said.

The researcher explained that the main advantage of GSE-based algorithms over many other well-established motion planning algorithms lies in the significant improvement of computational efficiency.

This naturally leads to strong applicability of GSE-based algorithms in applications where planning is time-sensitive.

“Instead of using computationally heavy dedicated collision detection modules, these algorithms take advantage of the novel notion of ‘normalized size’, which gives the maximum representation of free space that is accessible from a point in the environment, approximately It is tantamount to updating the human perception of a ‘safe’ place to move around,” said Adhvaith Rajkumar.

This, in essence, significantly improves the search speed of the environment by requiring only very few iterations of the GSE-based algorithm to connect the initial and target regions.

Explaining the applications of “motion planning” algorithms, Satdal Ghosh, assistant professor in the Department of Aerospace Engineering, IIT Madras, said, “Drones equipped with our algorithms can be of major use during disaster management and response scenarios.”

“In the wake of a disaster event such as an earthquake, UAVs are often deployed to survey affected areas and scan debris for search and rescue missions. Since such applications can advance UAV paths in a time-critical manner. need to be employed, our algorithms can play an important role,” he said.

“Broadly speaking, the class of GSE-based algorithms has promising potential in autonomous applications such as warehouse material movement, inspection of project commissioning, drone delivery, disaster management, self-driving cars, and the like. Coordinated motion in a multiple strategy For in-vehicle set-up also these algorithms can be availed,” he said.

The current state of this research, according to the team, is limited to the theoretical development and improvement of GSE-based algorithms and broadly realistic simulation-based validation of the same.

The researchers plan to apply these algorithms to unmanned aerial and ground vehicles in the near future.

“In dynamic environments, where knowledge of the environment is limited to information from on-board sensors or when mission commands interfere with the movement of vehicles due to dynamically evolving mission-critical requirements, for example intelligence, surveillance and reconnaissance (ISR)” ) operations or planetary exploration using rovers usually called for frequent time-critical re-planning of motion,” Ghosh said.

“Even in such cases, our current study suggests that due to the unique nature of the visual fields computed by these algorithms at different points in the environment, motion planning on the go by our GSE-based algorithms It becomes quite easy,” Ghosh said.