Geometric constraint solving with conics and linkages

https://doi.org/10.1016/S0010-4485(01)00114-2Get rights and content

Abstract

Most geometric construction methods of geometric constraint solving systems use line and circle (rule and compass) as basic drawing tools. In this paper, by introducing conics and linkages, we provide a set of complete drawing tools for the construction approach of geometric constraint solving. Using these tools, we may enlarge the drawing scope of the construction approach and still keep the elegant style of geometric solutions to geometric constraint solving. As applications, we obtain pure geometric solutions to three sets of well-known constraint problems: the 10 Apollonian drawing problems, the 13 cases of a smallest tri-connected constraint graph, and constraint problems with distance constraints only.

Introduction

Geometric constraint solving (GCS) is the central topic in much of the current work of developing intelligent and parametric CAD systems. It also has applications in chemical molecular modeling, linkage design, computer vision and computer aided instruction [8]. GCS algorithms accept the declarative description of geometric diagrams or engineering drawings as input and output a drawing procedure for the diagram. There are four main approaches to GCS: the graph analysis approach [19], [20], [26], [29], the rule-based approach [2], [9], [14], [23], [27], [30], [31], the numerical computation approach [18], [25], [28], and the symbolic computation approach [4], [10], [24]. In practice, combinations of these approaches are often used to obtain the best result. The work in Ref. [15] is such a successful attempt.

The graph analysis and the rule-based approaches are also called geometric construction approach. In this approach, a pre-treatment is carried out to transform the constraint problem into a constructive form that is easy to draw. In most cases, we construct the diagram sequentially with ruler and compass. This approach follows the tradition of engineering drawings, where ruler and compass are used as basic drawing tools. The construction approaches are the first choices for most GCS systems because they may provide complete, fast and stable solutions. However, the drawing scope of ruler and compass are restrictive. It is well known that using ruler and compass alone, we can draw diagrams that can be described by a sequence of triangular equations (see Section 2.1) of degree less than or equal to two. As shown in Ref. [10], we may find very simple design problems that cannot be solved with ruler and compass. The aim of this paper is to introduce new drawing tools to enlarge the drawing scope of the construction approach and still keep its elegant solution style.

We will consider two kinds of tools: conics and linkages. We prove that the diagrams that can be drawn with conics are those that can be described by a sequence of triangular equations of degree less than or equal to four. Therefore, the drawing scope by introducing conics is essentially larger than using lines and circles only. Since equations with degree ≤4 can be solved explicitly by radicals, the computation procedure is still complete and efficient as in the case of ruler and compass construction. Furthermore, for some diagrams that can be drawn with ruler and compass, we may give much simpler construction procedures by using conics. As an example, simple solutions are obtained for the 10 Apollonian drawing problems.

Linkages are one of the most basic kinds of mechanisms. Here, we will use them as drawing tools. Based on a result of Kempe, we prove that linkages as tools are complete in a certain sense. We also give an algorithm to find linkages in a constrained diagram. As a consequence, we prove that all constraint problems involving distance constraints only can be solved sequentially with linkages.

To solve the equations raised from linkage constructions, besides the often used numerical and symbolic computation methods, we introduce a geometric method which is to use the linkages to generate loci and then to find the intersection of these loci by searching the points on the loci. This geometric method is based on dynamic generation of geometric loci which are widely used in dynamic geometry software [12].

As an application, we show that a simplest constraint graph which is beyond the scope of Hoffmann and Owen's triangle decomposition methods can be transformed to pure geometric constructive form if linkages are allowed as construction tools. The linkages used in the construction are three kinds of four-bar linkages. One of the four-bar linkages is the traditional one. The other two kinds of four-bar linkages are newly introduced.

The rest of this paper is organized as follows. In Section 2, we show how to solve constraint problems with conics. In Section 3, we introduce linkages as a drawing tool. In Section 4, we present the solution to the smallest tri-connected constraint graph. In Section 5, we present the conclusion.

Section snippets

Constructive sequence with drawing tools

We will introduce several basic concepts for drawing geometric diagrams with general drawing tools. In this paper, a 2D drawing tool is a mechanism with one degree of freedom. The locus of a drawing tool is the trace of a geometric object in the mechanism, usually a point or a line. When the trace object is a point, the locus is a plane curve. A diagram can be drawn with a set of drawing tools sequentially if the geometric objects in the diagram can be listed in an orderP1,P2,…,Pmsuch that each

Scope of construction with linkages

In this paper, by a linkage, we mean a mechanism with one degree of freedom and consisting of links with fixed lengths and rotation joins only [17]. One linkage is shown in Fig. 5. In the linkage, points A and B are fixed and when point D rotates on circle CIR(B,|DB|), point E generates a locus which is represented by the dotted curve.

Since we can draw lines and conics with linkages [8], drawing with ruler, compass and conics is a special case of drawing with linkages. So, it is natural to ask

Solution to a smallest tri-connected constraint graph

Hoffman and Owen's triangle decomposition methods are among the fundamental methods of GCS. The basic idea is to divide the constraint graph into three parts, construct the each part separately, and then assemble them together. The difference is that Hoffman uses a bottom-up approach and Owen uses a top-down approach. Constraint graphs that can be solved by these methods are non-tri-connected graphs [19], [29]. A simplest constraint graph that cannot be solved with these methods is the one in

Conclusion

Geometric construction approaches are usually among the first choices by GCS systems because these approaches may provide complete and efficient solutions to geometric constraint problems. In this paper, we extend the scope of construction approaches by introducing two new kinds of drawing tools: conics and linkages. Using these tools, construction approaches become a complete method for all kinds of constraint problems, theoretically. Practically, many interesting constraint problems have been

Acknowledgements

The first author would like to thank Professors Shang-Ching Chou and Jian-Xin Ge for valuable discussions. This work is supported in part by Chinese NSF under an outstanding youth grant (No. 69725002) and by a National Key Basic Research Project of China (No. G1998030600).

Xiao-Shan Gao received his PhD from the Chinese Academy of Sciences in 1988. Since 1996, he has been a Professor in the Institute of Systems Science, Chinese Academy of Sciences. He has published over 60 research papers and two monographs. His research interests include: automated reasoning, symbolic computation, intelligent CAD and CAI (computer aided instruction).

References (32)

  • S.C. Chou

    Mechanical geometry theorem proving

    (1987)
  • S.C. Chou et al.

    Mechanical formula derivation in elementary geometries

    Proc ISSAC-90

    (1990)
  • X.S. Gao

    Automated geometry diagram construction and engineering geometry

  • Gao XS, Yang WQ. Simulation of complex linkages, Preprint, Institute of Systems Science, Academia Sinica,...
  • X.S. Gao et al.

    Geometry expert

    (1998)
  • X.S. Gao et al.

    Automated generation of Kempe linkage and its complexity

    J Computer Science and Technology

    (1999)
  • Cited by (9)

    View all citing articles on Scopus

    Xiao-Shan Gao received his PhD from the Chinese Academy of Sciences in 1988. Since 1996, he has been a Professor in the Institute of Systems Science, Chinese Academy of Sciences. He has published over 60 research papers and two monographs. His research interests include: automated reasoning, symbolic computation, intelligent CAD and CAI (computer aided instruction).

    Kun Jiang received his BS from HeiLongJiang University in 1994 and his MS from Harbin Institute of Technology in 1996. Now he is a PhD candidate in the Institute of Systems Science, Chinese Academy of Sciences. His research interests are geometric constraint solving and intelligent CAD.

    Chang-Cai Zhu received his BS in mathematics from University of Science and Technology of China in 1997. Now he is a PhD candidate in the Institute of Systems Science, Chinese Academy of Sciences. His research interests are design and analysis of linkage and geometric constraint solving.

    View full text