Exact and efficient construction of Minkowski sums of convex polyhedra with applications

We present an exact implementation of an efficient algorithm that computes Minkowski sums of convex polyhedra in R³. Our implementation is complete in the sense that it does not assume general position. Namely, it can handle degenerate input, and it produces exact results. We also present applications of the Minkowski-sum computation to answer collision and proximity queries about the relative placement of two convex polyhedra in R³. The algorithms use a dual representation of convex polyhedra, and their implementation is mainly based on the Arrangement package of Cgal, the Computational Geometry Algorithm Library. We compare our Minkowski-sum construction with the only three other methods that produce exact results we are aware of. One is a simple approach that computes the convex hull of the pairwise sums of vertices of two convex polyhedra. The second is based on Nef polyhedra embedded on the sphere, and the third is an output-sensitive approach based on linear programming. Our method is significantly faster. The results of experimentation with a broad family of convex polyhedra are reported. The relevant programs, source code, data sets, and documentation are available at http://www.cs.tau.ac.il/~efif/CD and a short movie [Fogel E, Halperin D. Video: Exact Minkowski sums of convex polyhedra. In: Proceedings of 21st annual ACM symposium on computational geometry. 2005. p. 382-3] that describes some of the concepts portrayed in this paper can be downloaded from http://www.cs.tau.ac.il/~efif/CD/Mink3d.avi.     

Shortest Path Geometric Rounding

Exact implementations of algorithms of computational geometry are subject to exponential growth in running time and space. In particular, coordinate bit-complexity can grow exponentially when algorithms are cascaded : the output of one algorithm becomes the input to the next. Cascading is a significant problem in practice. We propose a geometric rounding technique: shortest path rounding . Shortest path rounding trades accuracy for space and time and eliminates the exponential cost introduced by cascading. It can be applied to all algorithms which operate on planar polygonal regions, for example, set operations, transformations, convex hull, triangulation, and Minkowski sum. Unlike other geometric rounding techniques, shortest path rounding can round vertices to arbitrary lattices, even in polar coordinates, as long as the rounding cells are connected. (Other rounding techniques can only round to the integer grid.) On the integer grid, shortest path rounding introduces less combinatorial change and geometric error than the other rounding methods. Three algorithms are given for shortest path rounding, one of which we have used in industrial application software since 1992. In combination with recent advances in exact floating point evaluation of numerical primitives, shortest path geometric rounding yields a practical solution to numerical issues in computational geometry. Geometric algorithms can be implemented exactly on floating point input coordinates; the exact output coordinates can be rounded to accurate floating point approximations; and the cost of each arithmetic operation is only a little more than if it were implemented as a single hardware floating point operation. Received February 7, 1997; revised September 9, 1998.     

Exact Minkowksi Sums of Polyhedra and Exact and Efficient Decomposition of Polyhedra into Convex Pieces

We present the first exact and robust implementation of the 3D Minkowski sum of two non-convex polyhedra. Our implementation decomposes the two polyhedra into convex pieces, performs pairwise Minkowski sums on the convex pieces, and constructs their union. We achieve exactness and the handling of all degeneracies by building upon 3D Nef polyhedra as provided by Cgal. The implementation also supports open and closed polyhedra. This allows the handling of degenerate scenarios like the tight passage problem in robot motion planning. The bottleneck of our approach is the union step. We address efficiency by optimizing this step by two means: we implement an efficient decomposition that yields a small number of convex pieces, and develop, test and optimize multiple strategies for uniting the partial sums by consecutive binary union operations. The decomposition that we implemented as part of the Minkowski sum is interesting in its own right. It is the first robust implementation of a decomposition of polyhedra into convex pieces that yields at most O(r ²) pieces, where r is the number of edges whose adjacent facets comprise an angle of more than 180 degrees with respect to the interior of the polyhedron. This work was partially supported by the Netherlands’ Organisation for Scientific Research (NWO) under project no. 639.023.301.     

Controlled linear perturbation

We present an algorithmic solution to the robustness problem in computational geometry, called controlled linear perturbation, and demonstrate it on Minkowski sums of polyhedra. The robustness problem is how to implement real RAM algorithms accurately and efficiently using computer arithmetic. Approximate computation in floating point arithmetic is efficient but can assign incorrect signs to geometric predicates, which can cause combinatorial errors in the algorithm output. We make approximate computation accurate by performing small input perturbations, which we compute using differential calculus. This strategy supports fast, accurate Minkowski sum computation. The only prior robust implementation uses a less efficient algorithm, requires exact algebraic computation, and is far slower based on our extensive testing.     

Algorithm to calculate the Minkowski sums of 3-polytopes based on normal fans

Prompted by the development of algorithms for analysing geometric tolerancing, this article describes a method to determine the Minkowski sum for 3-dimensional polytopes. This method is based exclusively on intersection operations on normal cones, using the properties of the normal fan of a Minkowski sum obtained by common refinement of the normal fans of the operands. It can be used to determine from which vertices of the operands the vertices of the Minkowski sum derive. It is also possible to determine to which facets of the operands each facet of the Minkowski sum is oriented. The basic properties of the algorithms can be applied to n-polytopes.First, the main properties of the duality of normal cones and primal cones associated with the vertices of a polytope are described. Next, the properties of normal fans are applied to define the vertices and facets of the Minkowski sum of two polytopes. An algorithm is proposed, which generalises the method. Lastly, there is a discussion of the features of this algorithm, developed using the OpenCascade environment.     

Improvements to algorithms for computing the Minkowski sum of 3-polytopes

A Minkowski sum is a geometric operation that is equivalent either to the vector additions of all points in two operands or to the sweeping of one operand around the profile of the other without changing the relative orientation. Applications of Minkowski sums are found in computer graphics, robotics, spatial planning, and CAD. This paper presents two algorithms for computing Minkowski sum of convex polyhedron in three space (3-polytopes). Both algorithms are improvements on current ones found in the literature. One is based on convex hulls and the other on slope diagrams. The original convex hull based Minkowski algorithm is costly, while the original slope diagram based algorithms require the operation of stereographic projection from 3D to 2D for merging the slope diagrams of the two operands. Implementation of stereographic projection is complicated which increases the computation time and reduces the accuracy of the geometric information that is needed for constructing the resultant solid. This paper reports on improvements that have been made to these two algorithms and their implementation. These improvements include using vector operations to find the interrelations between points, arcs and regions on a unit sphere for the slope diagram algorithm, and addition of a pre-sorting procedure before constructing convex hull for convex hull based Minkowski sum algorithm. With these improvements, the computation time and complexity for both algorithms have been reduced significantly, and the computational accuracy of the slope diagram algorithm has been improved. This paper also compares these two algorithms to each other and to their original counterparts. The potential for extending these algorithms to higher dimensions is briefly discussed.     

Fast and robust retrieval of Minkowski sums of rotating convex polyhedra in 3-space

We present a novel method for fast retrieval of exact Minkowski sums of pairs of convex polytopes in R³, where one of the polytopes frequently rotates. The algorithm is based on pre-computing a so-called criticality map, which records the changes in the underlying graph structure of the Minkowski sum of the two polytopes, while one of them rotates. We give tight combinatorial bounds on the complexity of the criticality map when the rotating polytope rotates about one, two, or three axes. The criticality map can be rather large already for rotations about one axis, even for summand polytopes with a moderate number of vertices each. We therefore focus on the restricted case of rotations about a single, though arbitrary, axis.Our work targets applications that require exact collision detection such as motion planning with narrow corridors and assembly maintenance where high accuracy is required. Our implementation handles all degeneracies and produces exact results. It efficiently handles the algebra of exact rotations about an arbitrary axis in R³, and it well balances between preprocessing time and space on the one hand, and query time on the other.We use Cgal arrangements and in particular the support for spherical Gaussian maps to efficiently compute the exact Minkowski sum of two polytopes. We conducted several experiments (i) to verify the correctness of the algorithm and its implementation, and (ii) to compare its efficiency with an alternative (static) exact method. The results are reported.     

Dynamic Minkowski sums under scaling

In many common real-world and virtual environments, there are a significant number of repeated objects, primarily varying in size. Similarly, in many complex machines, there are a significant number of parts which also vary in size rather than shape. This repetition saves in both design and production costs. Recent research in robotics has also shown that exploiting workspace repetition can significantly increase efficiency. In this paper, we address the need to support computation reuse in fundamental operations. To this end, we propose an algorithm to reuse the computation of the Minkowski sum when an object is transformed by uniform scaling. The Minkowski sum is a fundamental operation in many areas of robotics, such as motion planning, computer vision, and mathematical morphology, and has been studied extensively over the last four decades. We present two methods for dynamically updating Minkowski sums under scaling, the first of which updates the sum under uniform scaling of arbitrary polygons and polyhedra, and the second of which updates the sum under non-uniform scaling of convex polyhedra. Ours are the first methods that study the Minkowski sum under this type of transformation. Our results show speed gains between one and two orders of magnitude over recomputing the Minkowski sum from scratch for each repeated object, and we discuss applications for motion planning, CAD, rapid prototyping, and shape decomposition.     

Determining the directional contact range of two convex polyhedra

The directional contact range of two convex polyhedra is the range of positions that one of the polyhedra may locate in along a given straight line so that the two polyhedra are in collision. Using the contact range, one can quickly classify the positions along a line for a polyhedron as “safe” for free of collision with another polyhedron, or “unsafe” otherwise. This kind of contact detection between two objects is important in CAD, computer graphics and robotics applications. In this paper we propose a robust and efficient computation scheme to determine the directional contact range of two polyhedra. We consider the problem in its dual equivalence by studying the Minkowski difference of the two polyhedra under a duality transformation. The algorithm requires the construction of only a subset of the faces of the Minkowski difference, and resolves the directional range efficiently. It also computes the contact configurations when the boundaries of the polyhedra are in contact.     

Mathematical morphological operations of boundary-represented geometric objects

The resemblance between the integer number system with multiplication and division and the system of convex objects with Minkowski addition and decomposition is really striking. The resemblance also indicates a computational technique which unifies the two Minkowski operations as a single operation. To view multiplication and division as a single operation, it became necessary to extend the integer number system to the rational number system. The unification of the two Minkowski operations also requires that the ordinary convex object domain must be appended by a notion of inverse objects or negative objects. More interestingly, the concept of negative objects permits further unification. A nonconvex object may be viewed as a mixture of ordinary convex object and negative object, and thereby, makes it possible to adopt exactly the same computational technique for convex as well as nonconvex objects. The unified technique, we show, can be easily understood and implemented if the input polygons and polyhedra are represented by their slope diagram representations.     

Rational bases for convex polyhedra

In Wachspress (1975) [2] rational bases were constructed for convex polyhedra whose vertices were all of order three. The restriction to order three was first removed by Warren (1996) [3] and his analysis was refined subsequently by Warren and Schaefer (2004) [4]. A new algorithm (GADJ) for finding the denominator polynomial common to all the basis functions was exposed in Dasgupta and Wachspress (2007) [1] for convex polyhedra with all vertices of order three. This algorithm is applied here for generating bases for general convex polyhedra.     

Finding the intersection of two convex polyhedra

Given two convex polyhedra in three-dimensional space, we develop an algorithm to (i) test whether their intersection is empty, and (ii) if so to find a separating plane, while (iii) if not to find a point in the intersection and explicitly construct their intersection polyhedron. The algorithm runs in timeO (nlogn), where n is the sum of the numbers of vertices of the two polyhedra. The part of the algorithm concerned with (iii) (constructing the intersection) is based upon the fact that if a point in the intersection is known, then the entire intersection is obtained from the convex hull of suitable geometric duals of the two polyhedra taken with respect to this point.     

AGRELs and BIPs: Metamorphosis as a Bezier curve in the space of polyhedra

The metamorphosis between two user-specified objects offers an intuitive metaphor for designing animations of deforming shapes. We present a new technique for interactively editing such deformations and for animating them in realtime. Besides the starting and ending shapes, our approach offers easy to use additional control over the deformations. The new Bezier Interpolating Polyhedron (BIP) provides a graphics representation of such a deforming object formulated mathematically as a point describing a Bezier curve in the space of all polyhedra. We replace, in the Bezier formulation, the traditional control points by arbitrary polyhedra and the vector addition by the Minkowski sum. BIPs are composed of Animated GRaphic ELement (AGRELs), which are faces with constant orientation, but with parametrized vertices represented by Bezier curves. AGRELs were designed to efficiently support smooth realtime animation on commercially available rendering hardware. We provide a tested algorithm for automatically computing BIPs from the sequence of control polyhedra and demonstrate its applications to animation design.     

An exact algorithm with learning for the graph coloring problem

Given an undirected graph G=(V,E), the Graph Coloring Problem (GCP) consists in assigning a color to each vertex of the graph G in such a way that any two adjacent vertices are assigned different colors, and the number of different colors used is minimized. State-of-the-art algorithms generally deal with the explicit constraints in GCP: any two adjacent vertices should be assigned different colors, but do not specially deal with the implicit constraints between non-adjacent vertices implied by the explicit constraints. In this paper, we propose an exact algorithm with learning for GCP which exploits the implicit constraints using propositional logic. Our algorithm is compared with several exact algorithms among the best in the literature. The experimental results show that our algorithm outperforms other algorithms on many instances. Specifically, our algorithm allows to close the open DIMACS instance 4-Fullins_5.     

An Efficient Method for Computing Exact State Space of Petri Nets With Stopwatches

In this paper, we address the issue of the formal verification of real-time systems in the context of a preemptive scheduling policy. We propose an algorithm which computes the state-space of the system, modeled as a time Petri net with stopwatches, exactly and efficiently, by the use of Difference Bounds Matrices (DBM) whenever possible and automatically switching to more time and memory consuming general (convex) polyhedra only when required. We propose a necessary and sufficient condition for the need of general polyhedra. We give experimental results comparing our implementation of the method to a full DBM over-approximation and to an exact computation with only general polyhedra.     

Space sweep solves intersection of convex polyhedra

Summary Plane-sweep algorithms form a fairly general approach to two-dimensional problems of computational geometry. No corresponding general space-sweep algorithms for geometric problems in 3- space are known. We derive concepts for such space-sweep algorithms that yield an efficient solution to the problem of solving any set operation (union, intersection, ...) of two convex polyhedra. Our solution matches the best known time bound of O(n log n), where n is the combined number of vertices of the two polyhedra.     

Exact Geodesics and Shortest Paths on Polyhedral Surfaces

We present two algorithms for computing distances along convex and non-convex polyhedral surfaces. The first algorithm computes exact minimal-geodesic distances and the second algorithm combines these distances to compute exact shortest-path distances along the surface. Both algorithms have been extended to compute the exact minimal-geodesic paths and shortest paths. These algorithms have been implemented and validated on surfaces for which the correct solutions are known, in order to verify the accuracy and to measure the run-time performance, which is cubic or less for each algorithm. The exact-distance computations carried out by these algorithms are feasible for large-scale surfaces containing tens of thousands of vertices, and are a necessary component of near-isometric surface flattening methods that accurately transform curved manifolds into flat representations.     

A memetic algorithm for the Minimum Sum Coloring Problem

Given an undirected graph G, the Minimum Sum Coloring Problem (MSCP) is to find a legal assignment of colors (represented by natural numbers) to each vertex of G such that the total sum of the colors assigned to the vertices is minimized. This paper presents a memetic algorithm for MSCP based on a tabu search procedure with two neighborhoods and a multi-parent crossover operator. Experiments on a set of 77 well-known DIMACS and COLOR 2002–2004 benchmark instances show that the proposed algorithm achieves highly competitive results in comparison with five state-of-the-art algorithms. In particular, the proposed algorithm can improve the best known results for 15 instances.     

Convex recolorings of strings and trees: Definitions, hardness results and algorithms

A coloring of a tree is convex if the vertices that pertain to any color induce a connected subtree; a partial coloring (which assigns colors to some of the vertices) is convex if it can be completed to a convex (total) coloring. Convex colorings of trees arise in areas such as phylogenetics, linguistics, etc., e.g., a perfect phylogenetic tree is one in which the states of each character induce a convex coloring of the tree.When a coloring of a tree is not convex, it is desirable to know “how far” it is from a convex one, and what are the convex colorings which are “closest” to it. In this paper we study a natural definition of this distance—the recoloring distance, which is the minimal number of color changes at the vertices needed to make the coloring convex. We show that finding this distance is NP-hard even for a colored string (a path), and for some other interesting variants of the problem. In the positive side, we present algorithms for computing the recoloring distance under some natural generalizations of this concept: the first generalization is the uniform weighted model, where each vertex has a weight which is the cost of changing its color. The other is the non-uniform model, in which the cost of coloring a vertex v by a color d is an arbitrary non-negative number cost(v,d). Our first algorithms find optimal convex recolorings of strings and bounded degree trees under the non-uniform model in time which, for any fixed number of colors, is linear in the input size. Next we improve these algorithm for the uniform model to run in time which is linear in the input size for a fixed number of bad colors, which are colors which violate convexity in some natural sense. Finally, we generalize the above result to hold for trees of unbounded degree.