Current Research work

1. Guidance and Control of Missiles
Modern missions require the guidance algorithm to achieve terminal and path constrained engagement with minimum control effort. Guidance formulation using optimal control theory can achieve these requirements.

In our lab, optimal guidance algorithms are designed to meet the terminal and path constraints. The technique used to develop these guidance algorithms is Model Predictive Static Programming (MPSP): MPSP is inspired from the predictive nature of Model Predictive Control and takes the advantage of computational efficiency of Approximate Dynamic programming. MPSP satisfies the terminal constraints by correcting the control history at each grid point iteratively by a close form update formula.

A revised version of MPSP is also formulated where the control variable is represented as weighted function of some basis functions. This variety is known as ‘Quasi-Spectral MPSP’ (QS-MPSP). It reduces the optimization dimension and thereby reduces computation time of MPSP.  Path constraints are satisfied by another variety of Quasi-Spectral MPSP known as ‘Constrained QS-MPSP’. In this formulation the path constraints (look angle and lateral acceleration) are presented as inequality constraints on control variable only.  The solution is obtained by minimizing a quadratic cost subjected to terminal equality and path inequality constraints using standard nonlinear optimization techniques.

2. Deciphering Control Architecture of Voluntary Movements Using Movement Variability
The motor system is corrupted by noise which can affect the system at various levels from the sensory stage to the execution of movements. This inherent noise in the system is reflected in the behaviour as the movement variability. Here, we use the pattern of variability to better understand the nature of control of biological movements. In this work, the saccadic system has been used to study and understand variability in order to obtain better insight into the control of movements. Saccades are rapid eye movements that increase the acuity of vision by bringing the target image near to the fovea in the eye. The inter-trial variability in the saccade trajectory is studied in conjunction with a stochastic saccade generation model for understanding the control of saccadic eye movements. The work suggests that understanding variability in movements is a powerful tool to investigate movement control. Thus studying how biological systems produce accurate movements in the presence of noise may help engineer artificial systems that are as effective as human systems in producing flexible and accurate movements.

3. Integrated Guidance and Control of a Lunar Module for Optimal Soft-landing
The Lander spacecraft is expected to perform soft landing at the selected site on the lunar surface. The maneuver needs to be done with minimum fuel requirement because of space and weight limitations. The powered descent phase is divided into three operation regimes, namely (i) Rough Braking Phase (ii) Precision Braking Phase and (iii) Terminal Descent Phase.

In addition, the algorithm needs to have the capability to re-designate the landing to an alternate site in the precision phase in case any hazard is detected at the designated landing site. Also the terminal orientation of the lander needs to be ensured vertical with respect to local ground.

Additional requirements include that the algorithm should be computationally very efficient to be amenable for onboard implementation and it should preferably have sufficient algorithmic generality to be extensible to future missions. The algorithms to be developed will be based on nonlinear optimal and nonlinear control theory and is expected to lead to substantial amount of fuel saving with enough robustness. Moreover, the general philosophy will be quite useful for other similar missions.

4. Nature Inspired Multiagent Multitarget Interception
Simultaneous Multiagent Multitarget (MA-MT) Interception Problem has been of interest for the past two decades in connection to missile salvo problem, navigation of mobile robots, reconnaissance problem etc. The current problem entails that a group of mobile agents will communicate among themselves in an unknown environment to search and explore a search space for locating target(s), which happen to be larger than the attacking agents. So on locating a target, an agent will approach and simultaneously signal neighboring agents to approach the target themselves, as observed in pack hunting. Until a critical number of agents have come in the vicinity of the target, the agents which have already “captured” the target will continue to circumnavigate the target from a “radius of attack” so as to restrict the motion of the target and thwart a retaliatory action from the target. When the critical number of agents have reached the radius of attack, the agents will attack the target together. So the overall interception is the ensemble of three aspects, viz. a “search phase”, “approach phase”, and “attack phase”.

Some possible solutions exist for missile interception problems which are based on Lyapunov Theory, variants of Proportional-Navigation (PN) guidance algorithms etc. Multiagent robotics navigation and associated problems like Coverage Control etc. have commended several classical approaches based on Proximity Functions, Navigation Functions etc. But many of them will require all-to-all communication, which is impractical for large number of agents in an unknown environment.  No single guidance and control strategy is going to achieve the three phases. Hence, inspiration from nature is sought to emulate the foraging behavior of wolf packs, insects like ants, bees etc., the flocking behavior of birds, fish shoals etc. In our current body of research, the objective is not to mimic the exact respective foraging behaviors which are complex processes and often largely empirical. Rather cues and signals that lead to successful foraging in nature we will be identified and incorporate those functionalities in the robots for successful interception in unknown environment.

Preliminary position based switched control law have been proposed by extending pack hunting behavior of wolves (originally proposed by Escobedo et al) to the MA-MT scenario. A combined velocity and position based behavioral control law based on local interaction and vision cone (originally proposed by Borzi et al) has been proposed which works in realistic predator-prey scenarios. Current efforts are underway to develop identical laws based on topological distance. Another direction of research underway is to frame the MA-MT problem as a minimization problem and solve as a Dynamic Optimization Problem. Hybrid variants of Evolutionary Algorithms are being explored to be suitable candidates.

5. Robust Adaptive Flight Control Design of Air-breathing Hypersonic Vehicles
The scramjet-powered air-breathing hypersonic vehicle is designed to fly at hypersonic speed (i.e. more than five times the speed of sound). It offers several distinct performance advantages over rocket-based systems for space access vehicles such as (1) no necessity of carrying oxidizer for propulsion (rather oxygen from atmosphere is used as the oxidizer), (2) no moving parts in the engine such as compressors and turbines, (3) very less time to reach the target/destination due to high speed etc. However, these performance advantages are dependent upon advances in current state-of-the art technologies in many critical areas, one of which happens to be the control system design.

The main objective of this research is to develop a “state-constrained robust nonlinear adaptive controller” for hypersonic vehicles to constrain the states within the prescribed bounds, and also minimize the possibility of inlet unstart during cruise phase of flight. The developed controller should ensure the stability of the vehicle in presence of modelling inaccuracy of the vehicle dynamics, thereby minimizing the necessity to have a very high fidelity model. The key idea be to quickly learn the model discrepancy function online using a fast disturbance observer and then utilize the updated model in state-constrained nonlinear control theory framework

6. Air Traffic Modelling and Optimal Management
Air traffic is increasing day-by-day. As aircrafts fly from airport to airport directly at high speed and they cannot halt on way, we need to design a trajectory for the flight starting from take-off to landing of aircraft. The trajectories needs to be designed carefully to avoid collisions, no-fly zones. Multiple aircrafts have different performance capabilities (speed, agility etc) and they are flying at a same time. This makes this problem interesting and complex.

The objective of research to design an algorithm which designs an optimal trajectory for aircraft in consensus with all other aircraft flying at same time. The delays in arrival, departure at airport affects the airport utility which may causes changes in many flight plans.

Ref :

7. Adaptive Control and Guidance of High Performance Aircrafts and UAVs
His current research interest are  Development of  L1 adaptive control and guidance algorithms for auto landing of UAVs with external disturbances, Development of Neuro-adaptive control and guidance with state estimation for Autolanding of UAVs with reactive collision avoidance, Development of  Image processing techniques with  Extended Kalman filter for runway / Road identification for auto landing applications of UAV.

UAVs are ubiquitous now a days. UAVs are being used in civilian, commercial and military applications. UAVs fly at low altitude and are susceptible to collision with other UAVs flying at similar altitude range. The main objective of this research is to develop a control and guidance algorithm with image processing and state estimation to sense and avoid collision with obstacle UAV while performing auto landing operations.

Figure : Autolanding of UAV with reactive collision Avoidance

8. Guidance and Control for Exo-atmospheric Kill Vehicles
For a hit-to-kill interceptor such as the Terminal High Altitude Area Defense (THAAD) missile system, effective terminal phase guidance and control is critical in ensuring zero miss distance, which in turn ensures warhead effectiveness. The interception of ballistic targets at high altitudes poses a significant challenge to guidance and control design due to:

  • Extremely high closing velocities
  • Very small time-to-go (time to intercept target).

Low dynamic pressures at high altitudes results in the lack of effectiveness of aerodynamic control surfaces

Fig 2: Exo-atmospheric missile engagement [1]

The missile however, has constraints on maneuverability due to:

  • Thrust constraints on the Divert and Attitude Control System (DACS).
  • Limited seeker field of view (FoV)

A guidance and control law is required to ensure that the missile detects the target throughout the engagement and ultimately carries out a successful intercept

Fig 2: THAAD Anti-Ballistic Missile system [2]

Conventional guidance and control techniques use individual loops to run guidance, control and estimation computations and have proved to be useful but introduce time lags between various loops. For high speed missiles in the terminal phase, where engagement times are in the order of a few seconds, a small increase in time lag can make all the difference between mission success and failure.

The project seeks to develop innovative, robust and effective guidance and control architectures for an Exo-atmospheric Kill Vehicle (EKV), that can run on-board the missile, Such as: Optimal control based guidance and control laws, integrated guidance and control (IGC) algorithms,  …etc.

Image References:



9. Artificial Pancreas For Type-1 Diabetes Patients Of India
Type 1 Diabetes Mellitus (T1DM) is a condition characterized by the body’s inability to produce insulin due to autoimmune destruction of the pancreatic beta cells. In the absence of insulin the glucose in the blood cannot be converted to energy by the cells in the body. Hence, T1DM patients depend upon exogenous insulin for regulating their blood glucose levels.

At present, a majority of the T1DM patients manually inject exogenous insulin by monitoring their glucose concentrations throughout the day. Our project aims at developing an Artificial Pancreas System in order to automate the exogenous insulin infusion process. The closed loop system would use the glucose concentrations obtained from continuous glucose monitoring sensors as input to calculate the required insulin dosage.

These estimated physiological parameters for Indian subjects will help customize the glucose-insulin model for Indian population. The mathematical model customized for Indian population will then be used for developing a glucose control feedback loop to determine the insulin dosage required. Advanced control strategies such as variants of model predictive control which predicts the glucose trajectory in the future for given insulin input. The mathematical model does not exactly replicates the glucose insulin dynamics as in human body, therefore there are many unaccounted variations which needs to be taken care of, and this will be done using neuro adaptive control. Finally arriving at most suitable algorithm for artificial pancreas system.

The control algorithm will then be embedded on a cost effective insulin pump that is being indigenously developed at IISc. The final closed loop system will comprise of a CGMS that wirelessly transmits glucose reading to the control algorithm embedded on the insulin pump.