Wm W » ° ' A° &* z 4. 1 * ^* **0< .( F *0, 'o . » * A ^ J ^ G v ^5^ a v V **T^-' .** r .0 *■* >. C», *o . » « A ** v % ^. v ^^ A," v-» - ■ - < j •- * .' **' * V V vv •J* b ,% ^ ** •YSfe*. U a* ^Wa* ^ <£ •fife"- ^ ^ ^Va- ^ c* ''fife*- ** ^ «*» A * :- ** « •■ * > V N .!••- %, a l" ^ a* A& •fi^'- % ** ^ ?7t % aO" 1 V^^T'*^' S>*.«iStf* ^ ^ )* >iJ^L'* ^> ' * O N '^^ •• *o A* . k ' » « <^A CT a • • • ♦ "*Q A*^ . V ' . „ <^. -1 o *° ** X Figure 5.-Two triangles formed by representing position of each target as vertex of triangle, and separation between lasers as length of baseline. and the various target separations as the baselines of the triangles (fig. 7). These triangles share some common angles and sides, and provide the following three equations with three unknown values. By using the law of sines, *E •TE sin (Oj) sin (0 F - 6 E ) and l E a ED sin (9 2 ) sin (6 E - 6 D ) By using the sum of angles law 180 = (e F - g d ) + e x + e 2 . By equating first and second equations, a relationship between Gj and 6 2 is found. The values of these angles can then be determined by substituting in the third equa- tion. The following two equations result: tan -l sin K, .(K x / K 2 ) + cos K^ where and *1 sin (9 F - Gg) K, d ED ^2 sin (9 E - 9 D ) K 3 = 180 - e F - e D . 9 2 = tan -1 sin K, _(K 2 / KJ + cos % The sides of the triangles (sensor-to-target distances) can then be calculated by using the law of sines, as shown in the following equations: 'FD Figure 6. -Angular position of three targets (D,E,F) on continuous miner from one Lasemet unit (L3) on MCS. and -D sin (6p - 9 D ) dr~ = d ED U E sin (9 E ■ e D ) dr^ " d FD Up sin (9 F - e D ) x sin 9 1 » x sin 9 2> x sin 9 2 As shown in the first routine, the position of each target relative to L3 can be determined by projecting the sides of the triangles onto the x and y axis. The heading (h) of the e F -e 8 f -8 E Figure 7.-Three triangles formed by representing Lasernet unit's position as common vertex, and targets separations as baselines of triangles. targets relative to the centerline of the cut can then be calculated from two target positions. These two routines demonstrate the minimum angular position information necessary to determine position and heading. Either the angular position of two same targets from two Lasernet units or the angular position of three targets relative to one Lasernet unit is necessary. How- ever, the current system is configured to support as many as three Lasernet units and five targets. The accuracy and reliability of this system are increased as more Lasernet units and more targets are added to the system. If any Lasernet unit fails, or target goes out of view, other lasers and targets will continue to provide adequate guidance information, thus increasing the reliability of the system. Additionally, at times when more than the minimum infor- mation is available, a sensor fusion algorithm which fuses all the gathered data together can provide a more accurate determination of position and heading. SENSOR FUSION ALGORITHM The current system (three lasers, five targets) will most often provide a redundant source of position and heading information. The accuracy of the position and heading determination can be increased by assigning weights (W) to the values obtained from the laser system and averaging the combined information. Each of the position and heading values [V(l),V(2),V(3)...V(n)] can be assigned a confidence level [C(l),C(2),C(3)...C(n)]. The confidence levels represent the accuracy and reliability of the laser system and are determined from prior experiments and analysis. These experiments and analysis will indicate times and conditions under which the data are accurate and reliable. Experiments are currently being conducted to determine the inherent accuracy of Lasernet. These tests will deter- mine Lasernet's angular position accuracy with respect to the target's range and angular position. An irregularity in the scanning motor is one possible source of inherent error. Aside from the inaccuracies inherent in the Lasernet unit, other analysis will provide information leading to the determination of confidence levels. It will determine rel- evant information such as the relationship between the angular separation between targets and the accuracy of the angular report. Also, it will determine the effect of certain geometric orientations of targets and lasers that affect angular accuracy. For example, a line of targets approach- ing a perpendicular orientation relative to the laser radial will result in a less reliable calculation of position and heading, and targets approaching a parallel orientation will result in an indeterminate situation. Additionally, environ- mental factors that affect accuracies of Lasernet's angular report, such as dust or water, will be evaluated. All this information will be used to predict the confidence of Lasernet's data. Once all experiments and analyses are complete, the expected performance of Lasernet will be known. The confidence of one laser unit may be determined from the data from other units. This situation may arise when the data from one or a few laser units can be used to deter- mine that another laser unit is operating outside its accu- rate range. The confidence level can then be lowered according to the results of the prior experiments. In another case, the confidence will be affected by a dust sensor on the machine that indicates a high concentration of dust. Again, the results of the dust performance tests will be used to predict the confidence of the data. Once a confidence level is determined for the data, average weights (W) can be calculated from the percent- age confidence levels. W(i) = C(i) / [C(l) + C(2) + ... + C(n)] The fused value V(f) of position and heading is then determined by V(f) = W(l) * V(l) + W(2) * V(2) + ... + W(n) * V(n) This value is the position and heading of the targets on the continuous miner relative to Lasernet units on the MCS. The target positions on the continuous miner and the Lasernet positions and orientations on the MCS are known; therefore, the relative position and heading of the continuous miner and MCS can be determined. At times when the continuous miner is using the MCS as a refer- ence, the orientation of the MCS with respect to the desired cut is known, therefore the continuous miner's position and heading relative to the desired cut can be determined. At times when the MCS is guiding from the targets on the continuous miner, the continuous miner's position and heading with respect to the desired cut are known, and the position and heading of the MCS with respect to the desired cut can be determined. CONCLUSIONS A method for position and heading determination to provide feedback for guidance of a continuous mining machine through the maneuvers required at the face area of a mine was described. It incorporated a commercially available angular position sensing system with a computer interface developed and tested at the Bureau. The current system consists of up to three laser-based optical scanning units which report the angular location of five retroreflective targets within their 90° field-of-view. The computer interface gathers the data from the Lasernet units, converts them into usable units, and employs two trigonometric routines to calculate the position and head- ing of the continuous miner. Experiments are currently being conducted to determine the expected performance of Lasernet. A sensor fusion algorithm that will use these data has been outlined and will be incorporated into the interface programming to increase the accuracy and reliability of the system. * U.S. GOVERNMENT PRINTING OFFICE: 611-012/00,082 INT.BU.OF MINES.PGH..PA 28920 ■o m £Q > i- CO c z m ?8 2 IT 5-m - 3 8 b z Ss oo o o (0 • c (0 O "O ft (A S 3" (D m O c > i— O "D "0 o 3 m ■o I— O < m 3D H ^3 90 r «» <$ O. ••n' A, u y - AT *> "••• <"p "A »* O 'bK °^^-'V V*^\/ %^-V . V-*^\/ %w^->° .. V^^ **** ,*>** w &*«* ►* •111!* v *.r« 9 v ^ % % *♦ .*J ^ t «° <$> * o « o ' ■$> »•- ^o^ •- ^oV* y o >*\.^&/V 0^.^1^>o >*..ii&. V C * :- ^o^ : ^oV* °i« ••!•' A. v y\ : -sk- : . X $$•• ./x --ft . X . l w ,/\3®^ ^ *...._.. -* /..i^i.A y..i^:..V .c°*..^t.°o ./ •"" ^'- *fe V* 0< v "« A^" "*v> *' * N. MANCHESTER I W» • &%^* . ^ ^ "^JR*^* «7 ** • r^^* . * * -rWBirt ^ ^» •