Apparatus for rapidly determining the trigonometric functions of an input angle

• US 4,077,063 A
• Filed: 08/11/1976
• Issued: 02/28/1978
• Est. Priority Date: 08/11/1976
• Status: Expired due to Term

• Current Assignee: Hughes Aircraft Company (Rtx Corporation)
• Original Assignee: Singer PLC (Hayleys Plc)
• Inventors: Lind, Thomas Peter
• Primary Examiner(s): Malzahn, David H.
• Application Number:
  US05/713,562
• Time in Patent Office: 566 Days
• Field of Search: 235/152, 235/197, 235/186
• US Class Current: 708/440
• CPC Class Codes: G06F 1/0356 by using two or more smaller tables, e.g. addressed by parts of the argument
• CPC Class Codes: G06F 2101/04 Trigonometric functions

Abstract

High-speed hardware is employed to:

(1) subdivide an input angle into a series of smaller subangles according to the sine approximation algorithm:

SIN(INPUT) ≅ SIN(A) + COS(A)SIN(B) + COS(A)SIN(C) + COS(A)SIN(D);

(2) process the above terms separately through individual look-up tables; and

(3) sum the table results to provide the sine function of the input angle.

The above algorithm is derived from a series of approximations relating to the sine and cosine functions of very small angles. The relative sizes of the above subangles is expressed by:

A >> B > C > D

the actual size of each subangle is dependent on the number of input bits, the size of the look-up tables available, and the accuracy desired.

Claims

1. Apparatus responsive to a signal representation of an input angle (INPUT) having Nin binary places for providing a signal representation of an approximation of the function SIN(INPUT) having Nout bits, comprising:

input means for receiving the Nin binary places of INPUT and identifying a set of at least three progressively smaller subangles A, B, C . . . LAST-1, and LAST formed by consecutive groups of sequential binary places such that;

space="preserve" listing-type="equation">INPUT = A + B + C + . . . LAST-1, + LAST,

and

space="preserve" listing-type="equation">INPUT >

A >

B >

C . . . LAST-1, >

LAST;

a set of function generator means responsive to the input means, each of which provides a term of the approximate sine relationship;

space="preserve" listing-type="equation">SIN(INPUT) ≅

SIN(A) + COS(A)SIN(B) + COS(A)SIN(C) + . . . + COS(A)SIN(LAST-1) + COS(A)SIN(LAST);

and

adder means responsive to each function generator for combining the terms therefrom to form the approximate sine relationship.

2. The apparatus of claim 1, wherein:

subangle A is the most significant subangle and is formed by the group of most significant bits (Na);

Subangle B is the next most significant subangle and is formed by the group of next most significant (Nb);

Subangle . . . ;

subangle LAST-1 is the next to least significant subangle and is formed by the group of next to least significant bits (Nlast -

     1) and

Subangle LAST is the least significant subangle and is formed by the group of least significant bits (Nlast).

3. Apparatus responsive to a signal representation of an input angle (INPUT) having Nin binary places for providing a signal representation of an approximation of the function SIN (INPUT) having Nout bits, comprising:

input means for receiving the Nin binary places of INPUT and identifying a set of at least three progressively smaller subangles A, B, C . . . LAST - 1, and LAST formed by consecutive groups of sequential binary places such that;

space="preserve" listing-type="equation">INPUT = A + B + C + . . . LAST - 1, + LAST,

space="preserve" listing-type="equation">input >

a >

b >

c . . . last - 1,>

last,

space="preserve" listing-type="equation">a = (the MSB of A) + (the remaining LSB of A) = Ap + As;

and

a set of function generator means responsive to the input means, each of which provides a term of the approximate sine relationship;

space="preserve" listing-type="equation">SIN(INPUT ≅

SIN(Ap + As) + COS(Apb + Asb) SIN(B) + COS(Apc + Asc) SIN(C) + . . . + COS(Aplast - 1 + Aslast -

     1) SIN(LAST -

     1) + COS(Aplast + Aslast) SIN(LAST)

in which the Ap_- portion for subangles B through LAST is any group of the most significant bits of A, and the As_- portion for subangles B through LAST is at least the most significant bits of the remaining least significant bits of A; and

adder means responsive to each function generator for combining the terms therefrom to the approximate sine relationship.

4. Apparatus responsive to a signal representation of an input angle (INPUT) having Nin binary places for providing a signal representation of an approximation of the function SIN(INPUT) having Nout bits, comprising:

input means for receiving the Nin binary places of INPUT and identifying a set of at least three progressively smaller subangles A, B, C . . . LAST - 1, and LAST formed by consecutive groups of sequential binary places such that;

space="preserve" listing-type="equation">INPUT = A + B + C + . . . LAST - 1 + LAST,

space="preserve" listing-type="equation">input a >

b >

c . . . last - 1 >

last,

space="preserve" listing-type="equation">a = (the MSB of A) + (the remaining LSB of A) = Ap + As;

and

a set of function generator means responsive to the input means, each of which provides a term of the approximate sine relationship;

space="preserve" listing-type="equation">SIN(INPUT) ≅

SIN(Ap + As) + COS(Apb + Kb)SIN(B) + COS(Apc + Kc)SIN(C) + . . . + COS(Aplast - 1 + Klast -

     1)SIN(Last -

     1) + COS(Aplast + Klast)SIN(Last)

in which the Ap_- portion for subangle B through LAST is any group of the most significant bits of A, and the K_- is a constant having a predetermined value; and

adder means responsive to each function generator for combining the terms therefrom to form the approximate sine relationship.

5. The apparatus of claim 4, wherein each predetermined constant value is equal to one of the end points of the range of values for the subangle associated therewith.

6. The apparatus of claim 4, wherein each predetermined constant value is equal to the maximum value of the range of values for the subangle associated therewith.

7. The apparatus of claim 4, wherein the approximate sine relationship is a monotonic relationship.

8. The apparatus of claim 4, wherein at least a portion of the function generator means in the set are memory look-up tables for the functions of the term associated therewith.

9. The apparatus of claim 8, wherein all of the function generator means in the set are ROMs forming a set of ROMs, ROM-A, ROM-B, ROM-C . . . ROM-Last - 1, and ROM-LAST, one ROM for each term of the approximate sine relationship.

10. The apparatus of claim 9, wherein:

Rom-a provides the first term SIN(A) and has two to the Na address locations;

Rom-b provides the second term COS(Apb + Asbmax)SIN(B) and has two to the (Napb + Nb) address locations;

Rom-c provides the third term COS(Apc + Ascmax)SIN(C) and has two to the (Napc + Nc) address locations;

. . .

Rom-last-1 provides the LAST - 1 term COS(Aplast - 1 + Aslast - 1max)SIN(LAST -

     1) and has two to the (Naplast - Nlast) address locations; and

Rom-last provides the LAST term COS(Aplast + Aslastmax) SIN(LAST) and has two to the (Naplast + Nlast) address locations.

11. The apparatus of claim 10, wherein the number of bit places in each of the ROMs progressively decreases according to the relationship:

space="preserve" listing-type="equation">ROM-A bits >

ROM-B bits >

. . . >

ROM-LAST bits

to eliminate leading zeros generated by the decreasing significance of the product terms therefrom.

12. The apparatus of claim 9, wherein the ROMs are PROMs.

13. The apparatus of claim 12, wherein the input means is hardwired to the appropriate ROMs to present the binary places of each subangle to the appropriate ROMs.

14. The apparatus of claim 13, wherein the hardwired ROm inputs are distributed so that:

space="preserve" listing-type="equation">Na >

>

Nb ≦

Nc ≦

Nd.