My current Research work focuses on 'trajectory optimization techniques in uncertain and dynamic environments'
The following are a list of the projects that have been handled by me:
1) Trajectory optimization for reactive planning in dynamic environments
The two primary contributions of the proposed trajectory optimization are (1): A computationally efficient method for computing the intersection space of collision avoidance constraints of large number of predicted obstacle trajectories.(2): An optimization framework to connect the current state to the solution space in time optimal fashion.This work is published in IROS-2014.
The following are a list of the projects that have been handled by me:
1) Trajectory optimization for reactive planning in dynamic environments
The two primary contributions of the proposed trajectory optimization are (1): A computationally efficient method for computing the intersection space of collision avoidance constraints of large number of predicted obstacle trajectories.(2): An optimization framework to connect the current state to the solution space in time optimal fashion.This work is published in IROS-2014.
|
|
2) Closed form characterization of collision free velocities and confidence bounds for non-holonomic robots in uncertain dynamic environments:
In this work, a probabilistic version of time scaled collision cone was obtained by representing obstacle states through generic probability distributions.The novelty of the current work, is on the reformulation of probabilistic constraints into deterministic algebraic constraints, . A small demonstration of the proposed method, along with a technical report is shown is shown below. This work is accepted in IROS-15.
|
|
3) Mobile Robot navigation amidst humans with intents and uncertainties
This work is developed for navigation of non holonomic robots in human centered environments, wherein the trajectories that a human can take are modeled as intentions, with distinct probability values assigned to it.An optimization framework is developed that achieves an elegant balance between the objective of minimizing risk and ease of avoidance maneuver . A small demonstration of the video is shown below. This work is accepted in CDC-15
This work is developed for navigation of non holonomic robots in human centered environments, wherein the trajectories that a human can take are modeled as intentions, with distinct probability values assigned to it.An optimization framework is developed that achieves an elegant balance between the objective of minimizing risk and ease of avoidance maneuver . A small demonstration of the video is shown below. This work is accepted in CDC-15
|
|
4) PRVO: Probabilistic Reciprocal Velocity Obstacle:
In this work a probabilistic version of the famous Reciprocal Velocity Obstacle is developed to incorporate state and estimation uncertainties, This work was novel , as a closed form characterization of the space of collision free velocities was developed taking into consideration state and actuation uncertainties.This work was published in IROS17.
|
|
5) Risk Aware Merging:
The main objective was to develop a risk aware merging behavior, for a traffic like scenario.This work is published in IV 2018
The main objective was to develop a risk aware merging behavior, for a traffic like scenario.This work is published in IV 2018
|
|
6) Probabilistic obstacle avoidance and object following: An overlap of Gaussians approach : Roman 2019:
We propose an optimization framework that solves for a sequence of controls with a guaranteed risk bound. We characterize probabilistic collision avoidance as a measure of overlap between two Gaussian distributions. These distributions are primarily inferred from the uncertainties of the robot and obstacle. Furthermore, we provide closed form expressions that can characterize this overlap as a function of the control input, and hence provide a solution which seamlessly integrates with the well known framework of MPC (Model predictive control).
ro-man_final.pdf | |
File Size: | 1983 kb |
File Type: |