Elsevier

Computer-Aided Design

Volume 35, Issue 14, December 2003, Pages 1337-1344
Computer-Aided Design

Isophote interpolation

https://doi.org/10.1016/S0010-4485(03)00063-0Get rights and content

Abstract

The isophote is an important class of characteristic curves on a parametric surface. Accordingly, as a kind of feature based path generation method, isophote based tool paths has been proposed as an important path pattern for parametric surface machining. In this paper, an approach has been proposed for the isophote interpolation on a parametric surface. The paper has related the arc-length derivatives to time derivatives of parameters along the isophote. The results are then used to derive the parametric interpolation. The isophote curve interpolation is developed based on parametric interpolation. The proposed interpolation guarantees that interpolated points always stay on the parametric surface. An improvement interpolation has been presented to alleviate inclination errors and point deviation from the isophote. Simulations of isophote interpolation have been carried out to verify the effectiveness of the proposed algorithm. The proposed algorithm has applications in real time tool path interpolation for the machining of parametric surfaces.

Introduction

Computer numerical control (CNC) machining is an important method for machining complex three-dimensional part surfaces. Surface machining involves four main phases, namely, optimization of machining process parameters, tool path planning, path interpolation and servo control. Path interpolation is concerned of converting tool paths obtained from a tool path planning system into time-dependent commands for driving the servo control system of a CNC machine.

Traditionally, linear and circular interpolators are the most common interpolators for NC systems, because of a lower computational requirement that was important in the past when computers were slower. A well-known approach for developing these interpolators is the digital differential analyzer (DDA) approach that is based on mathematical integration [9], [11]. Qin [16] also developed a DDA algorithm for the interpolation of three-dimensional circular paths. Another interpolation approach is based on pattern recognition [4], [15]. In this approach, a set of rules is employed to determine the next step from a possible set of moves. The above-mentioned approaches were originally conceived with the assumption of hardware implementation of the interpolator. The advent of computing technologies has been driving a migration of the implementation of interpolators from hardware to software.

While linear and circular interpolators are only built-in software interpolators for a CNC system, the approximation scheme is utilized for CNC surface machining. According to the approximation scheme, once the machining tool paths are generated, line approximation scheme are used to discretize the tool paths into line segments for the generation of NC codes. Real time operation of the CNC machine uses the downloaded NC codes as input. Further, arc approximation schemes were proposed to improve the smoothness of machined surfaces. Qiu [17] proposed to use circular arc interpolation for tool path generation. Curve fitting with arc splines was suggested by Yeung [29]. Biarcs were used to approximate complex tool paths in which line and arc segments were special cases of biarcs [22].

It was then recognized that those interpolation approaches were insufficient for the machining of complex surfaces. Disadvantages of the approximation approach include significant contour and orientation errors, heavy loading of the communication and large memory requirement for the CNC controller and considerable acceleration and deceleration for the line segment executions [10], [13], [19].

Today, computing speed for CNC controllers has increased dramatically and research on free-form curve interpolators has become quite active. Consequently, a real time path approach [10], [12], [18] is proposed as an alternative to off-line path approximation schemes. This aims to equip the CNC system with built-in interpolators for complex curves. In the real time path approach, the CNC system with built-in free-form curve interpolators inputs curve geometry parameters directly and generates motion commands in real time. Accordingly machining accuracy and product quality can be greatly improved while machining time is also significantly reduced.

Parametric curve is the most important form of tool paths for surface machining. Considerable attention was paid to the interpolation of parametric curves. General geometric properties of parametric curves were related to kinematics and coordinated motion of the machine axes in [2], [3]. Zhang [30] implemented a real time non-uniform rational B-spline interpolator for a multi-axis robot. Wang [23] proposed a real time quintic spline interpolator. Koren [10] introduced the concept of surface interpolators in which the tool orientations along a path are obtained from the parametric surface model. There were also efforts to improve the feedrate accuracy [14] and to use variable feedrate for path interpolation. Farouki [5], [6] proposed and implemented real time interpolators for Pythagorean-hodograph (PH) curves that are two-dimensional curves. A constant feed and reduced angular acceleration interpolation algorithm is proposed for parametric curve in multi-axis machining [7]. Yeh [27] also proposed a parametric curve interpolator with speed control and further presented an interpolation method with confined chord error [28].

Implicit curve is another form of tool paths for surface machining. While much of the earlier work in real time curve interpolations assume the use of parametric curves, recently there has been an increase in attention to real time interpolations for implicit curves [14], [19], [20], [24], [25], [26].

Drive surface methods is an important approach for tool path generation [1]. According to drive surface methods, tool paths can be generated by intersecting a part surface with a family of drive surfaces such as planes and cylinders. But in general, the tool path obtained by intersecting a part surface with a drive surface is a hybrid curve [21]. The hybrid curve is a novel form of tool paths, which is different from parametric curves and implicit curves. Accordingly hybrid curve interpolation [21] is proposed with tool paths generated by intersecting a parametrically modeled part surface with a family of implicitly defined drive surfaces.

According to Ref. [8], the common weakness of the current tool path generation methods is feature-blindness. The usual strategy is to adopt a preset mathematical rule to generate paths for the entire surface. The geometry difference that may exist among the scattered features/geometry on a surface, however, cannot be taken into consideration. The isophote is an important class of characteristic curves on a parametric surface. Then, as a kind of feature based path topology, isophote based tool paths has been proposed as an important path pattern for parametric surface machining in Ref. [8] where a tracing method for the isophote is proposed based on grid subdivision for off-line path approximation for the machining of parametric surface.

This paper is focused on the isophote interpolation in real time for parametric surface machining. It is noted that an isophote can be represented as neither a parametric curve, nor an implicit curve, nor a hybrid curve. Therefore, existing interpolation methods are not applicable to isophote interpolation. An approach for the isophote interpolation is proposed in this paper. The strategy is to carry out the isophote interpolation based on parametric interpolation. The arc-length derivatives of the parameters are related to time derivatives of parameters with feedrate along the isophote in Section 2. Based on the results, the parametric interpolation is developed in Section 3. In Section 4, the interpolation and its improvement for isophote are developed. Details of the algorithm and examples are provided for isophote interpolation for parametric surfaces in Section 5.

Section snippets

Isophote

The isophote is an important class of characteristic curves on a surface. An isophote is defined as a set of points on the surface at which normal vectors form a preset constant angle with a given fixed reference vector. SupposeS(u,w)=[x(u,w)y(u,w)z(u,w)]for0≤u,w≤1represents a parametric surface. The normal of the surface at any point isN=Su×SwwhereSu=dxdudydudzdu,Sw=dxdwdydwdzdwAccordingly, the first order and second order derivatives of the normal with respect to parameters can be derived asNa

Parametric interpolation

In this section, derivatives of the parameters along the isophote with respect to the arc-length have been analyzed and then related to their derivatives with respect to time. The results are then used to derive the parametric interpolation algorithm.

Isophote interpolation

The interpolation of an isophote can be carried out through the parametric interpolation. Suppose on the isophote curve, the current interpolated vector is Pi=(xiyizi) with its corresponding parametric vector pi and the next interpolated vector Pi+1=(xi+1yi+1zi+1) with its corresponding parametric vector pi+1. Through Eq. (26), the parametric vector pi+1 can be computed based on pi. Then according to the parametric surface (1), the next interpolated point Pi+1 can be obtained,Pi+1=S(ui+1,wi+1)=[

Algorithm and examples

In terms of proposed approach above, the detailed algorithm for isophote interpolation is presented as Fig. 1. The algorithm starts with input of the surface parameters, reference vector and feedrate. At each parametric point, the arc-length derivatives of parameters are calculated. Using the derivatives, the next parametric point is determined. What follows is to evaluate the inclination error at the obtained parametric point. If the inclination error is over prescribed error limit, the

Conclusion

An approach has been proposed for the isophote interpolation. This enables interpolated points along an isophote to be generated in real time for CNC surface machining. The paper has analytically related the arc-length derivatives of the parameters to time derivatives of parameters with given feedrate along the isophote. The results are then used to develop the parametric interpolation. The isophote interpolation is directly based on the parametric interpolation. The proposed interpolation

Hai-Yin Xu is currently an associate professor at College of Computer Science and Technology, Huazhong University of Science and Technology (HUST), People's Republic of China. Previously he was a postdoctoral research associate at King's College, University of London from 2001. Prior to moving to London, he spent two years as a research assistant at City University of Hong Kong from 1998. He received BS and MS degrees in Control Engineering from Northwestern Polytechnical University and HUST,

References (30)

Cited by (8)

View all citing articles on Scopus

Hai-Yin Xu is currently an associate professor at College of Computer Science and Technology, Huazhong University of Science and Technology (HUST), People's Republic of China. Previously he was a postdoctoral research associate at King's College, University of London from 2001. Prior to moving to London, he spent two years as a research assistant at City University of Hong Kong from 1998. He received BS and MS degrees in Control Engineering from Northwestern Polytechnical University and HUST, and PhD degree in mechanical engineering from HUST. His research interests are in CNC and computer-aided design and manufacturing (CAD/CAM), computer graphics and image, virtual reality, and bioinformatics.

Hon-Yuen Tam received a BS degree from Georgia Institute of Technology, USA, a MS and a PhD (1990) degrees from Stanford University, USA, all in mechanical engineering. He is now a university lecturer of the department of manufacturing engineering and engineering management at City University of Hong Kong. His current research interests are in automated surface finishing, numerical control, flexible automation systems and surface inspection.

Jian J. Zhang is Professor of Computer Graphics at the National Centre for Computer Animation, Bournemouth University, United Kingdom. He has taken a number of academic positions both in the UK and in China. His current research interests include computer graphics, geometric modeling, computer animation and physics-based simulation.

View full text