Table of Contents Author Guidelines Submit a Manuscript
Journal of Applied Mathematics
Volume 2015 (2015), Article ID 827572, 7 pages
http://dx.doi.org/10.1155/2015/827572
Research Article

Function Synthesis Algorithm of RTD-Based Universal Threshold Logic Gate

1Hangzhou Institute of Service Engineering, Hangzhou Normal University, Hangzhou 311121, China
2Department of Information Science & Electronic Engineering, Zhejiang University, Hangzhou 310027, China

Received 19 March 2015; Revised 13 May 2015; Accepted 31 May 2015

Academic Editor: Georgios Sirakoulis

Copyright © 2015 Maoqun Yao et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The resonant tunneling device (RTD) has attracted much attention because of its unique negative differential resistance characteristic and its functional versatility and is more suitable for implementing the threshold logic gate. The universal logic gate has become an important unit circuit of digital circuit design because of its powerful logic function, while the threshold logic gate is a suitable unit to design the universal logic gate, but the function synthesis algorithm for the -variable logical function implemented by the RTD-based universal logic gate (UTLG) is relatively deficient. In this paper, three-variable threshold functions are divided into four categories; based on the Reed-Muller expansion, two categories of these are analyzed, and a new decomposition algorithm of the three-variable nonthreshold functions is proposed. The proposed algorithm is simple and the decomposition results can be obtained by looking up the decomposition table. Then, based on the Reed-Muller algebraic system, the arbitrary -variable function can be decomposed into three-variable functions, and a function synthesis algorithm for the -variable logical function implemented by UTLG and XOR2 is proposed, which is a simple programmable implementation.

1. Introduction

With the improvement in integrated circuit integration, the complementary metal oxide semiconductor (CMOS) technology is gradually approaching its physical limitations. The resonant tunneling device (RTD) has better performance and features, such as negative differential resistance characteristic, self-latching, high speed, and functional versatility [1, 2]. The universal logic gate, which has a powerful logic function, has become an important unit to implement -variable logical functions [3], and the RTD is more suitable for implementing the universal logic gate because of its negative differential resistance characteristic [46]. So, the RTD will probably become the main electronic device in the next generation of integrated circuits [7, 8].

Though the circuit of an -variable logical function implemented by the universal logic gate will be simpler, a different universal logic gate requires its corresponding synthesis algorithm to implement a function. Some function synthesis algorithms have been proposed in the literature [912], but these algorithms are not suitable for implementing an arbitrary -variable function by the RTD-based universal threshold logic gate (UTLG) [13]. And the algorithm [14] which can implement a three-variable nonthreshold function by UTLGs is relatively complicated, and the implemented circuit structure is also complicated.

In this paper, based on the Reed-Muller expansion, the three-variable nonthreshold functions are classified. Two categories of these are analyzed, and a new decomposition algorithm of the three-variable nonthreshold functions is proposed. Then a function synthesis algorithm which can implement an arbitrary -variable logical function by UTLGs is proposed. The proposed function synthesis algorithm provides a new scheme for designing integrated circuits by RTD devices.

2. Background

2.1. Threshold Logic

A threshold logic gate is defined as a logic gate with binary input variables and a single binary output. Its internal parameters are as follows: binary input variables,   , a set of integer weights,   , and a threshold and an output , such that its input-output relationship can be expressed as [15]

Formula (1) can also be presented as . If a logic function can be implemented with a single threshold logic gate, the function is called a threshold function; otherwise, it is called a nonthreshold function [15].

2.2. Spectral Technique

Spectral technique is a mathematical transformation method. It can convert binary data from the Boolean domain into the spectral domain by matrix transformations, and the information will not be lost [16]. In the spectral domain , for an -variable logical function , its input and output have kinds of states, and the truth vector is

The spectral-coefficient vector is given bywhere is a Rademacher-Walsh matrix.

As for a three-variable function, the spectral-coefficient vector is , where is a zero-order spectral-coefficient, and , , and are one-order, two-order, and three-order spectral coefficients, respectively.

2.3. Reed-Muller Expansion

A Reed-Muller expansion is a standard expansion in the AND/XOR algebraic system. A given function can be expressed as the XOR of basic entry; its coefficient is called the RM expansion coefficient [17]. Given an -variable function , its RM expansion coefficient vector (0-polarity; in this paper we only use 0-polarity expansion coefficient) iswhere , , is the Kronecker product, and is the truth vector of the function in the Boolean domain. As for a three-variable function , its RM expansion is

3. Decomposition Algorithm of Three-Variable Nonthreshold Functions

In this section, the three-variable nonthreshold functions are classified, and a new decomposition algorithm of the three-variable nonthreshold functions is proposed.

3.1. Determine the Three-Variable Threshold Function

The zero-order and one-order spectral-coefficient can determine whether a function is a threshold function [16]. The spectral-coefficient classification table of all the three-variable threshold functions is given in Table 1, and some conclusions as follows [16].

Table 1: Spectral-coefficient classification of all the three-variable threshold functions.

(1) The zero-order and one-order spectral-coefficients of the given function are calculated and arranged in numerically descending magnitude order, and this is the of the function; if the appear in Table 1, the function is a threshold function. The threshold value of a threshold function and weights can be obtained by in Table 1. The relationship between the and listed value is

The sign of is the same as . The threshold can be calculated as

(2) If the function is not a threshold function, the derived number will not appear in Table 1.

3.2. New Decomposition Algorithm of Three-Variable Nonthreshold Functions

According to the properties of the spectral coefficient, the three-variable nonthreshold function can be decomposed into three-variable threshold functions by a series of conversions [14], but for some three-variable nonthreshold functions, this algorithm [14] is relatively complex. So, we will propose a simpler algorithm which can decompose those three-variable nonthreshold functions into three-variable threshold functions.

For all the three-variable functions, there are 104 threshold functions and 152 nonthreshold functions [16].

First, according to Table 1, the 104 threshold functions are divided into three categories [16].(1)Among the absolute values of all the zero-order and one-order spectral-coefficients, there is one number equal to 8 and three numbers equal to 0.(2)Among the absolute values of all the zero- and one-order spectral-coefficients, there is one number equal to 6 and three numbers equal to 2.(3)Among the absolute values of all the zero-order and one-order spectral-coefficients, there is one number equal to 0 and three numbers equal to 4.

Then, the 152 nonthreshold functions can be divided into four categories.(1)Among the absolute values of all the spectral coefficients, the maximum number is from the two-order spectral coefficient.(2)Among the absolute values of all the spectral coefficients, the maximum number is from the three-order spectral coefficient.(3)Among the absolute values of all the zero-order and one-order spectral-coefficients, there is one number equal to 4.(4)Among the absolute values of all the zero-order and one-order spectral-coefficients, there are two numbers equal to 4.

The numbers of these four categories of three-variable nonthreshold functions are 54, 18, 33, and 47, respectively. If the function belongs to the second and third categories of the three-variable nonthreshold functions, the algorithm [14] which decomposes the function will be relatively complex. A nonthreshold function is decomposed into the threshold function by the conversion of and , so we propose a new decomposition algorithm of the three-variable nonthreshold functions based on the Reed-Muller algebraic system, and the decomposition process is as follows.(1)The RM expansion coefficient vectors of all the three-variable threshold functions are calculated as a reference table.(2)The RM expansion coefficient vector of the given function is calculated.(3)Search the reference table and find two threshold functions in which the XOR result of their RM expansion coefficient vectors just equals the calculated RM expansion coefficient vector in step 2.(4)The three-variable nonthreshold function can be expressed as the XOR of the found threshold functions of step 3.

After applying the process to the second and third categories of the three-variable nonthreshold functions, except two special functions, and , the others can be presented by the XOR of two threshold functions; that is, , where and are threshold functions. The decomposition results of the second and third categories of the three-variable nonthreshold functions are shown in Tables 2 and 3. In the tables, each vector of the columns is the output of the three-variable function for an 8-input combination from 000 to 111. Taking the first column of the first vector as an example in Table 2, it expresses the three-variable function   , where is the minterm of the function, and can be expressed as , where    and   , and they are the three-variable threshold functions denoted by the second and third columns of the first row in Table 2.

Table 2: The second category of the three-variable nonthreshold functions.
Table 3: The third category of the three-variable nonthreshold functions.    

4. The Synthesis Algorithm of -Variable Function Based on UTLG

The literature [13] proposed an RTD-based universal logic gate (UTLG) which can implement an arbitrary three-variable threshold function with a single UTLG. Figure 1 shows the schematic and symbol of UTLG; its input-output relationship can be expressed as

Figure 1: The universal threshold logic gate (UTLG), (a) schematic (b) symbol.

The first and fourth categories of the three-variable nonthreshold functions can be implemented by 3 UTLGs, and the second and third categories of the three-variable nonthreshold functions can be implemented by 7 UTLGs [13]. In our proposed decomposition algorithm of three-variable nonthreshold functions we introduced a bivariate XOR function, which cannot be implemented by a single UTLG, so according to the structure of MOBILE circuit [18], we design an RTD-based bivariate XOR gate (XOR2). Figure 2 shows the schematic and symbol of XOR2. Thus, the second and third categories of the three-variable nonthreshold functions can be implemented by 2 UTLGs and 1 XOR2, and it simplifies the circuit. Figure 3 shows the simulation results of XOR2 circuit by HSPICE; the parameters of RTD and HFET are the same as UTLG circuit [13], and as observed from Figure 3, the proposed XOR2 has the correct logic functionality.

Figure 2: The bivariate XOR gate (XOR2), (a) schematic (b) symbol.
Figure 3: Transient waveforms of the proposed XOR2.

Currently, there is no algorithm which can implement arbitrary -variable functions by the UTLG. Therefore, we proposed an -variable function synthesis algorithm based on UTLG.

First, we analyze the RM expansion of a four-variable function [17]; .

By observation, the last eight terms of the RM expansion are equal to the first eight terms of the RM expansion by multiplying , respectively, and the operations of terms in the RM expansion function are XOR, so the four-variable function can be expressed as , the variables of are , and , and the RM expansion coefficient vector of is equal to the first eight elements of the RM expansion coefficient vector of . The variables of are also , , and , and the RM expansion coefficient vector of is equal to the last eight elements of the RM expansion coefficient vector of .

For an -variable function, based on the Reed-Muller algebraic system [17], it can also be expressed as , so we propose the synthesis algorithm of -variable functions based on UTLG; its process is as follows.(1)Calculate the RM expansion coefficient vector of the given -variable function .(2)In the first decomposition, , the subfunction is the variable of the function except , and its RM expansion coefficient vector is equal to the first elements of ; the subfunction is also the variable of the function except , and its RM expansion coefficient vector is equal to the last elements of .(3)In the second decomposition, , , the subfunctions , , , and are the variables of functions except and , and their RM expansion coefficient vectors are equal to the first elements of , the last elements of , the first elements of , and the last elements of , respectively.(4)Repeat step 3, until the subfunctions are the three-variable functions.(5)Judging all the three-variable subfunctions, if the function is a threshold function, it can directly be implemented by 1 UTLG; if the function belongs to the first and fourth categories of the three-variable nonthreshold functions, it can be implemented by 3 UTLGs [14]; if the function belongs to the second and third categories of the three-variable nonthreshold functions, looking up the decomposition results from Tables 2 and 3, it can be implemented by 2 UTLGs and 1 XOR2; if the function is or , it can be implemented by 1 UTLG and 2 XOR2s.(6)The of each decomposition equation can be regarded as the bivariate function, which can be implemented by 1 UTLG. And the bivariate XOR operation of the decomposition equations can be implemented by 1 XOR2.

Example 1. Implement the four-variable function:The RM expansion coefficient vector of is , the can be decomposed as , is a three-variable function, and its RM expansion coefficient vector is equal to the first 8 elements of ; that is, , . is a three-variable function, and its RM expansion coefficient vector is equal to the last 8 elements of ; that is, , .
belongs to the second category of the three-variable nonthreshold functions; looking up Table 2, , , . belongs to the fourth category of the three-variable nonthreshold functions, and it can be decomposed as [14] , , . Then can be implemented by 6 UTLGs and 2 XOR2s. Figure 4 shows the UTLG and XOR2 implementation of this function.

Figure 4: UTLG and XOR2 implementation of Example 1.

Example 2. Implement the five-variable function:The RM expansion coefficient vector of is , can be decomposed as , is a four-variable function, its RM expansion coefficient vector is equal to the first 16 elements of , that is, , is a four-variable function, and its RM expansion coefficient vector is equal to the last 16 elements of , that is, .
Continue to decompose . , is a three-variable function, its RM expansion coefficient vector is equal to the first 8 elements of , that is, , is also a three-variable function, and its RM expansion coefficient vector is equal to the last 8 elements of , that is, .
Continue to decompose . , is a three-variable function, its RM expansion coefficient vector is equal to the first 8 elements of , that is, , is a three-variable function, and its RM expansion coefficient vector is equal to the last 8 elements of , that is, .
, , , and are the three-variable threshold functions, and can be implemented by 7 UTLGs and 3 XOR2s. Figure 5 shows the UTLG and XOR2 implementation of this function.
The proposed synthesis algorithm of the -variable function based on UTLG can decompose the arbitrary -variable function into three-variable functions, and all the three-variable nonthreshold functions are divided into four categories, and two categories of them can be implemented by UTLG and XOR2. Thus, arbitrary -variable functions can be implemented by UTLG and XOR2, and if the given function is decomposed into the threshold functions, the circuit structure will be simple.
Up to now, there is no algorithm which can implement arbitrary -variable function by UTLG. Reference [13] just proposed the algorithm which can implement arbitrary three-variable function. So, we provided new logic units UTLG and XOR2 and a new algorithm to implement RTD-based arbitrary -variable function.

Figure 5: UTLG and XOR2 implementation of Example 2.

5. Conclusion

In this paper, the 152 three-variable nonthreshold functions are divided into four categories, and a new decomposition algorithm of the three-variable nonthreshold functions is proposed. If the function belongs to the second and third categories of the three-variable nonthreshold functions, it can be implemented by 2 UTLGs and 1 XOR2, and it simplifies the implemented circuit structure. Based on the Reed-Muller algebraic system, the arbitrary -variable function can be decomposed into three-variable functions, and the function synthesis algorithm for the -variable function which can be implemented by UTLG and XOR2 is proposed; that is, the arbitrary -variable function can be implemented by UTLG and XOR2, and if the given function is decomposed into three-variable threshold functions, the circuit structure will be much simpler. The proposed logic units and algorithm present a new method to implement the arbitrary -variable function by RTD-based devices.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

This work was supported by the National Natural Science Foundation of China under Grant nos. 61271124 and 61471314 and Zhejiang Provincial Natural Science Foundation of China under Grant nos. LY13F010001 and LY15F010011.

References

  1. N. Muramatsu, H. Okazaki, and T. Waho, “A novel oscillation circuit using a resonant-tunneling diode,” in Proceedings of the IEEE International Symposium on Circuits and Systems (ISCAS '05), pp. 2341–2344, May 2005. View at Publisher · View at Google Scholar · View at Scopus
  2. P. Mazumder, S. Kulkarni, M. Bhattacharya, J. P. Sun, and G. I. Haddad, “Digital circuit applications of resonant tunneling devices,” Proceedings of the IEEE, vol. 86, no. 4, pp. 664–686, 1998. View at Publisher · View at Google Scholar · View at Scopus
  3. X. Chen and S. L. Hurst, “A consideration of the minimum number of input terminals on universal logic gates and their realization,” International Journal of Electronics, vol. 50, no. 1, pp. 1–13, 1981. View at Publisher · View at Google Scholar · View at MathSciNet · View at Scopus
  4. V. Beiu, J. M. Quintana, and M. J. Avedillo, “implementations of threshold logic—a comprehensive survey,” IEEE Transactions on Neural Networks, vol. 14, no. 5, pp. 1217–1243, 2003. View at Publisher · View at Google Scholar · View at Scopus
  5. Y. X. Zheng and C. Huang, “Complete logic functionality of reconfigurable RTD circuit elements,” IEEE Transactions on Nanotechnology, vol. 8, no. 5, pp. 631–642, 2009. View at Publisher · View at Google Scholar · View at Scopus
  6. S. M. Mirhoseini, M. J. Sharifi, and D. Bahrepour, “New RTD-based general threshold gate topologies and application to three-input XOR logic gates,” Journal of Electrical and Computer Engineering, vol. 35, no. 1, Article ID 463925, pp. 1–4, 2010. View at Publisher · View at Google Scholar · View at Scopus
  7. K. K. Likharev, “Hybrid CMOS/nanoelectronic circuits: opportunities and challenges,” Journal of Nanoelectronics and Optoelectronics, vol. 3, no. 3, pp. 203–230, 2008. View at Publisher · View at Google Scholar · View at Scopus
  8. J. Lee, S. Choi, and K. Yang, “A new low-power RTD-based 4:1 multiplexer IC using an InP RTD/HBT MMIC technoligy,” in Proceedings of the 22nd International Conference on Indium Phosphide and Related Materials, pp. 1–3, Kagawa, Japan, June 2010. View at Publisher · View at Google Scholar · View at Scopus
  9. S. Kolłodziński and E. Hrynkiewicz, “An utilisation of boolean differential calculus in variables partition calculation for decomposition of logic functions,” in Proceedings of the IEEE Symposium on Design and Diagnostics of Electronic Circuits & Systems (DDECS '09), pp. 34–37, April 2009. View at Publisher · View at Google Scholar · View at Scopus
  10. E. Hrynkiewicz and S. Kołodziński, “Non-disjoint decomposition of logic functions in Reed-Müller spectral domain,” in Proceedings of the 13th IEEE International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS '10), pp. 293–296, IEEE, Vienna, Austria, April 2010. View at Publisher · View at Google Scholar · View at Scopus
  11. E. Hrynkiewicz and S. Kołodziński, “An Ashenhurst disjoint and non-disjoint decomposition of logic functions in Reed-Muller spectral domain,” in Proceedings of the 17th International Conference on Mixed Design of Integrated Circuits and Systems (MIXDES '10), pp. 200–204, June 2010. View at Scopus
  12. F. Yu, L. F. Wang, R. H. Tan, and H. Jin, “An improved functional decomposition method based on FAST and the method of removal and operation,” in Proceedings of the International Conference on System Science and Engineering (ICSSE '12), pp. 487–492, Dalian, China, July 2012. View at Publisher · View at Google Scholar · View at Scopus
  13. Y. Wei and J. Z. Shen, “Novel universal threshold logic gate based on RTD and its application,” Microelectronics Journal, vol. 42, no. 6, pp. 851–854, 2011. View at Publisher · View at Google Scholar · View at Scopus
  14. R. Zhang, P. Gupta, L. Zhong, and N. K. Jha, “Threshold network synthesis and optimization and its application to nanotechnologies,” IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 24, no. 1, pp. 107–118, 2005. View at Publisher · View at Google Scholar · View at Scopus
  15. S. Muroga, Threshold Logic and Its Application, John Wiley & Sons, New York, NY, USA, 1971. View at MathSciNet
  16. S. L. Hurst, J. C. Muzio, and D. M. Miller, Spectral Techniques in Digital Logic, Academic Press, London, UK, 1985.
  17. K. L. Kodandapani and R. V. Setlur, “A note on minimal Reed-Muller canonical forms of switching functions,” IEEE Transactions on Computers, vol. 26, no. 3, pp. 310–313, 1977. View at Google Scholar · View at MathSciNet
  18. K. J. Chen and K. Maezawa, “Monostable-bistable transition logic elements (MOBILES) based on monolithic integration of resonant tunneling diodes and FETs,” Japanese Journal of Applied Physics, vol. 34, no. 2, pp. 1199–1203, 1995. View at Publisher · View at Google Scholar · View at Scopus