We consider the problem of sequential robotic manipulation of deformable objects using tools. Previous works have shown that differentiable physics simulators provide gradients to the environment state and help trajectory optimization to converge orders of magnitude faster than model-free reinforcement learning algorithms for deformable object manipulation. However, such gradient-based trajectory optimization typically requires access to the full simulator states and can only solve short-horizon, single-skill tasks due to local optima. In this work, we propose a novel framework, named DiffSkill, that uses a differentiable physics simulator for skill abstraction to solve long-horizon deformable object manipulation tasks from sensory observations. In particular, we first obtain short-horizon skills using individual tools from a gradient-based optimizer, using the full state information in a differentiable simulator; we then learn a neural skill abstractor from the demonstration trajectories which takes RGBD images as input. Finally, we plan over the skills by finding the intermediate goals and then solve long-horizon tasks. We show the advantages of our method in a new set of sequential deformable object manipulation tasks compared to previous reinforcement learning algorithms and compared to the trajectory optimizer.
more »
« less
This content will become publicly available on March 31, 2025
PolyARBerNN: A Neural Network Guided Solver and Optimizer for Bounded Polynomial Inequalities
Constraints solvers play a significant role in the analysis, synthesis, and formal verification of complex cyber-physical systems. In this article, we study the problem of designing a scalable constraints solver for an important class of constraints named polynomial constraint inequalities (also known as nonlinear real arithmetic theory). In this article, we introduce a solver named PolyARBerNN that uses convex polynomials as abstractions for highly nonlinears polynomials. Such abstractions were previously shown to be powerful to prune the search space and restrict the usage of sound and complete solvers to small search spaces. Compared with the previous efforts on using convex abstractions, PolyARBerNN provides three main contributions namely (i) a neural network guided abstraction refinement procedure that helps selecting the right abstraction out of a set of pre-defined abstractions, (ii) a Bernstein polynomial-based search space pruning mechanism that can be used to compute tight estimates of the polynomial maximum and minimum values which can be used as an additional abstraction of the polynomials, and (iii) an optimizer that transforms polynomial objective functions into polynomial constraints (on the gradient of the objective function) whose solutions are guaranteed to be close to the global optima. These enhancements together allowed the PolyARBerNN solver to solve complex instances and scales more favorably compared to the state-of-the-art nonlinear real arithmetic solvers while maintaining the soundness and completeness of the resulting solver. In particular, our test benches show that PolyARBerNN achieved 100X speedup compared with Z3 8.9, Yices 2.6, and PVS (a solver that uses Bernstein expansion to solve multivariate polynomial constraints) on a variety of standard test benches. Finally, we implemented an optimizer called PolyAROpt that uses PolyARBerNN to solve constrained polynomial optimization problems. Numerical results show that PolyAROpt is able to solve high-dimensional and high order polynomial optimization problems with higher speed compared to the built-in optimizer in the Z3 8.9 solver.
more »
« less
- Award ID(s):
- 2139781
- PAR ID:
- 10502340
- Publisher / Repository:
- Association for Computing MachineryNew YorkNYUnited States
- Date Published:
- Journal Name:
- ACM Transactions on Embedded Computing Systems
- Volume:
- 23
- Issue:
- 2
- ISSN:
- 1539-9087
- Page Range / eLocation ID:
- 1 to 26
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Fisman, D. ; Rosu, G. (Ed.)When augmented with a Pseudo-Boolean (PB) solver, a Boolean satisfiability (SAT) solver can apply apply powerful reasoning methods to determine when a set of parity or cardinality constraints, extracted from the clauses of the input formula, has no solution. By converting the intermediate constraints generated by the PB solver into ordered binary decision diagrams (BDDs), a proof-generating, BDD-based SAT solver can then produce a clausal proof that the input formula is unsatisfiable. Working together, the two solvers can generate proofs of unsatisfiability for problems that are intractable for other proof-generating SAT solvers. The PB solver can, at times, detect that the proof can exploit modular arithmetic to give smaller BDD representations and therefore shorter proofs.more » « less
-
Huisman, Marieke ; Pasareanu, Corina ; Zhan, Naijun (Ed.)This paper presents a formally verified quantifier elimination (QE) algorithm for first-order real arithmetic by linear and quadratic virtual substitution (VS) in Isabelle/HOL. The Tarski-Seidenberg theorem established that the first-order logic of real arithmetic is decidable by QE. However, in practice, QE algorithms are highly complicated and often combine multiple methods for performance. VS is a practically successful method for QE that targets formulas with low-degree polynomials. To our knowledge, this is the first work to formalize VS for quadratic real arithmetic including inequalities. The proofs necessitate various contributions to the existing multivariate polynomial libraries in Isabelle/HOL. Our framework is modularized and easily expandable (to facilitate integrating future optimizations), and could serve as a basis for developing practical general-purpose QE algorithms. Further, as our formalization is designed with practicality in mind, we export our development to SML and test the resulting code on 378 benchmarks from the literature, comparing to Redlog, Z3, Wolfram Engine, and SMT-RAT. This identified inconsistencies in some tools, underscoring the significance of a verified approach for the intricacies of real arithmetic.more » « less
-
The focus of this paper is on the development of an open loop controller for type 1 diabetic patients which is robust to meal and initial condition uncertainties in the presence of hypoand hyperglycemic constraints. Bernstein polynomials are used to parametrize the evolving uncertain blood-glucose. The unique bounding properties of these polynomials are then used to enforce the desired glycemic constraints. A convex optimization problem is posed in the perturbation space of the model and is solved repeatedly to sequentially converge on a sub-optimal solution. The proposed approach is demonstrated on the classic Bergman model for Type 1 diabetic patients.more » « less
-
more » « less
This paper proposes a novel method to efficiently solve infeasible low-dimensional linear programs (LDLPs) with billions of constraints and a small number of unknown variables, where all the constraints cannot be satisfied simultaneously. We focus on infeasible linear programs generated in the RLibm project for creating correctly rounded math libraries. Specifically, we are interested in generating a floating point solution that satisfies the maximum number of constraints. None of the existing methods can solve such large linear programs while producing floating point solutions.
We observe that the convex hull can serve as an intermediate representation (IR) for solving infeasible LDLPs using the geometric duality between linear programs and convex hulls. Specifically, some of the constraints that correspond to points on the convex hull are precisely those constraints that make the linear program infeasible. Our key idea is to split the entire set of constraints into two subsets using the convex hull IR: (a) a set X of feasible constraints and (b) a superset V of infeasible constraints. Using the special structure of the RLibm constraints and the presence of a method to check whether a system is feasible or not, we identify a superset of infeasible constraints by computing the convex hull in 2-dimensions. Subsequently, we identify the key constraints (i.e., basis constraints) in the set of feasible constraints X and use them to create a new linear program whose solution identifies the maximum set of constraints satisfiable in V while satisfying all the constraints in X. This new solver enabled us to improve the performance of the resulting RLibm polynomials while solving the corresponding linear programs significantly faster.
