More Like this

1. Using Deep Learning to Bootstrap Abstractions for Robot Planning

   Shah, Naman; Srivastava, Siddharth (May 2022, Proceedings of the International Conference on Autonomous Agents)

   This paper addresses the problem of learning abstractions that boost robot planning performance while providing strong guarantees of reliability. Although state-of-the-art hierarchical robot planning algorithms allow robots to efficiently compute long-horizon motion plans for achieving user desired tasks, these methods typically rely upon environment-dependent state and action abstractions that need to be hand-designed by experts. We present a new approach for bootstrapping the entire hierarchical planning process. This allows us to compute abstract states and actions for new environments automatically using the critical regions predicted by a deep neural network with an auto-generated robot-specific architecture. We show that the learned abstractions can be used with a novel multi-source bi-directional hierarchical robot planning algorithm that is sound and probabilistically complete. An extensive empirical evaluation on twenty different settings using holonomic and non-holonomic robots shows that (a) our learned abstractions provide the information necessary for efficient multi-source hierarchical planning; and that (b) this approach of learning, abstractions, and planning outperforms state-of-the-art baselines by nearly a factor of ten in terms of planning time on test environments not seen during training. 

   more » « less
2. Simultaneously Learning Transferable Symbols and Language Groundings from Perceptual Data for Instruction Following

   Gopalan, N; Rosen, E; Konidaris, G; Tellex, S (July 2020, Robotics: Science and Systems XVI)

   Enabling robots to learn tasks and follow instructions as easily as humans is important for many real-world robot applications. Previous approaches have applied machine learning to teach the mapping from language to low dimensional symbolic representations constructe by hand, using demonstration trajectories paired with accompanying instructions. These symbolic methods lead to data efficient learning. Other methods map language directly to high-dimensional control behavior, which requires less design effort but is data-intensive. We propose to first learning symbolic abstractions from demonstration data and then mapping language to those learned abstractions. These symbolic abstractions can be learned with significantly less data than end-to-end approaches, and support partial behavior specification via natural language since they permit planning using traditional planners. During training, our approach requires only a small number of demonstration trajectories paired with natural language—without the use of a simulator—and results in a representation capable of planning to fulfill natural language instructions specifying a goal or partial plan. We apply our approach to two domains, including a mobile manipulator, where a small number of demonstrations enable the robot to follow navigation commands like “Take left at the end of the hallway,” in environments it has not encountered before. 

   more » « less
3. Simultaneously Learning Transferable Symbols and Language Groundings from Perceptual Data for Instruction Following

   https://doi.org/10.15607/RSS.2020.XVI.102  

   Gopalan, N; Rosen, E; Konidaris, GD; Tellex, S (July 2020, Robotics: Science and Systems XVI)

   null (Ed.)

   Enabling robots to learn tasks and follow instructions as easily as humans is important for many real-world robot applications. Previous approaches have applied machine learning to teach the mapping from language to low dimensional symbolic representations constructed by hand, using demonstration trajectories paired with accompanying instructions. These symbolic methods lead to data efficient learning. Other methods map language directly to high-dimensional control behavior, which requires less design effort but is data-intensive. We propose to first learning symbolic abstractions from demonstration data and then mapping language to those learned abstractions. These symbolic abstractions can be learned with significantly less data than end-to-end approaches, and support partial behavior specification via natural language since they permit planning using traditional planners. During training, our approach requires only a small number of demonstration trajectories paired with natural language—without the use of a simulator—and results in a representation capable of planning to fulfill natural language instructions specifying a goal or partial plan. We apply our approach to two domains, including a mobile manipulator, where a small number of demonstrations enable the robot to follow navigation commands like “Take left at the end of the hallway,” in environments it has not encountered before. 

   more » « less
4. Socially Aware Navigation: A Non-linear Multi-objective Optimization Approach

   https://doi.org/10.1145/3453445  

   Banisetty, Santosh Balajee; Forer, Scott; Yliniemi, Logan; Nicolescu, Monica; Feil-Seifer, David (July 2021, ACM Transactions on Interactive Intelligent Systems)

   null (Ed.)

   Mobile robots are increasingly populating homes, hospitals, shopping malls, factory floors, and other human environments. Human society has social norms that people mutually accept; obeying these norms is an essential signal that someone is participating socially with respect to the rest of the population. For robots to be socially compatible with humans, it is crucial for robots to obey these social norms. In prior work, we demonstrated a Socially-Aware Navigation (SAN) planner, based on Pareto Concavity Elimination Transformation (PaCcET), in a hallway scenario, optimizing two objectives so the robot does not invade the personal space of people. This article extends our PaCcET-based SAN planner to multiple scenarios with more than two objectives. We modified the Robot Operating System’s (ROS) navigation stack to include PaCcET in the local planning task. We show that our approach can accommodate multiple Human-Robot Interaction (HRI) scenarios. Using the proposed approach, we achieved successful HRI in multiple scenarios such as hallway interactions, an art gallery, waiting in a queue, and interacting with a group. We implemented our method on a simulated PR2 robot in a 2D simulator (Stage) and a pioneer-3DX mobile robot in the real-world to validate all the scenarios. A comprehensive set of experiments shows that our approach can handle multiple interaction scenarios on both holonomic and non-holonomic robots; hence, it can be a viable option for a Unified Socially-Aware Navigation (USAN). 

   more » « less
5. Hierarchical Planning and Learning for Robots in Stochastic Settings Using Zero-Shot Option Invention

   https://doi.org/10.1609/aaai.v38i9.28903  

   Shah, Naman; Srivastava, Siddharth (March 2024, Proceedings of the AAAI Conference on Artificial Intelligence)

   This paper addresses the problem of inventing and using hierarchical representations for stochastic robot-planning problems. Rather than using hand-coded state or action representations as input, it presents new methods for learning how to create a high-level action representation for long-horizon, sparse reward robot planning problems in stochastic settings with unknown dynamics. After training, this system yields a robot-specific but environment independent planning system. Given new problem instances in unseen stochastic environments, it first creates zero-shot options (without any experience on the new environment) with dense pseudo-rewards and then uses them to solve the input problem in a hierarchical planning and refinement process. Theoretical results identify sufficient conditions for completeness of the presented approach. Extensive empirical analysis shows that even in settings that go beyond these sufficient conditions, this approach convincingly outperforms baselines by 2x in terms of solution time with orders of magnitude improvement in solution quality. 

   more » « less