Inter national J our nal of Robotics and A utomation (IJRA) V ol. 10, No. 4, December 2021, pp. 319 325 ISSN: 2089-4856, DOI: 10.11591/ijra.v10i4.pp319-325 319 The taxi function Alicia C. S ´ anchez Department of Mathematics, F aculty of Science, Uni v ersity of Extremadura, Badajoz, Spain Article Inf o Article history: Recei v ed Feb 3, 2021 Re vised May 16, 2021 Accepted Jul 23, 2021 K eyw ords: Automatic na vig ation systems Lane k eeping control Lateral control RAFU functions ABSTRA CT This paper in v estig ates the lane k eeping control and the lateral control of autonomous ground v ehicles, robots or the lik e considering the road agenc y formation unit (RAFU) functions. A strate gy based kno wing the real position of se v eral points of the trajec- tory is proposed to achie v e the lateral control purpose and maintain the lane k eeping errors within the prescribed performance boundaries. The RAFU functions are applied to achie v e these goals. The stability of these functions, their applica bility to approach an y arbitrary trajectory and the easy control of the possible error made on the approx- imation are useful adv antages in practice. This is an open access article under the CC BY -SA license . Corresponding A uthor: Alicia C. S ´ anchez Department of Mathematics, F aculty of Science Uni v ersity of Extremadura 06006 Badajoz, Spain Email: csanchezalicia@gmail.com 1. INTR ODUCTION Intelligent trans p or t systems (ITS) v ary in technologies applied: basi c traf c si g na l control syst ems, automatic number plate re cogn i tion, speed cameras to monitor security systems and the li k e. Mechatronics sys- tems (MS) include a combination of mechanical, electrical, telecommunications, control and computer science technologies. The robotics (R) in v olv es design, construction and use of robots and dra ws on the achie v ement of computer , mechanical or electronic engineering and mathematics. In particular ITS, MS and R include management metods such as automatic na vig ation systems, v e- hicle control and automatic dri ving. There are a multitude of papers about this topic. Here we cite some of them as e xample [1]-[5]. This w ork is conncerned about these topics, specically about lane k eeping control and lateral control. There is a wide of literature on current de v elopments in the eld of lateral and lane k eeping control of autonomous v ehicle motions [6]-[8]. In robotics, for e xample, the concept of lane-k eeping motion planning algorithms for mobile robots in order that the robot does not lea v e the lane if collision-free motion is a v ailable introduced in [9]. The road agenc y formation unit (RAFU) functions ha v e bee n studied i n Approximation Theory (the interested reader can see [10]-[17]). In this paper , the RAFU functions will be called the T axi functions. As when we use a taxi, the taxi dri v er tak e us to the destination, from kno wledge of the e xact posit ions of se v eral points of the road centerline and a width of the route, our main goal in this w ork is to obtain RAFU continuous functions that connect the initial point with t he nal point of a certain trajectory without touching the sides of the road. T o achie v e this aim, we only need that the robot has the necessary technology to kno w the e xact positions of some points of the route that it needs to follo w at an y time. In this w ork we are not concerned about what the machine must do in case of an y obstacle appears J ournal homepage: http://ijr a.iaescor e .com Evaluation Warning : The document was created with Spire.PDF for Python.
320 ISSN: 2089-4856 on the road. On this subject one can see for instance, [18]-[22]. Until no w , RAFU functions ha v e not been emplo yed for this purpose. But gi v en the ease and the accurac y with which these continuous functions approx- imate an y step f u nc tion and, therefore, an y continuous trajectory , we think that the use of these functions in ITS, MS and R could be useful, specically in lane k eeping and lateral control. Moreo v er , the easy control of the possible error made on the approximation of the e xact route is another adv antage in practice. In section 2 we solv e the lane k eeping and lat eral control problems in case of the routes are straight and perpendicular all of them. Section 3 is de v oted to solv e the same problems in case that the trajectory is a continuous function. The lane k eeping and lateral control problems in case of an arbitrary trajectory on the plane is studied in section 4 . Concluding remarks are in section 5 . This paper is illustrated with some e xampl es. 2. TRAJECT OR Y WHERE ALL R OUTES ARE STRAIGHT AND PERPENDICULAR Denition 1 Gi v en an arbitrary function f dened i n [ a, b ] and l et P s = { x 0 = a, x 1 , ..., x s = b } be a partition, we dene the RAFU Method on approximation to the function function f to all approximation procedure that uses functions C n dened in [ a, b ] to approach the function f where the functions C n ( x ) are dened by (1). C n ( x ) = f ( x 1 ) + s X i =2 [ f ( x i ) f ( x i 1 )] · F n,p ( x i 1 , x ) (1) being (2). F n,p ( x i 1 , x ) = 2 n p +1 x i 1 a + 2 n p +1 x x i 1 2 n p +1 p b x i 1 + 2 n p +1 x i 1 a (2) with p 1 a natural number . The functions C n ( x ) , n N are called RAFU Functions . Suppose we kno w that ( x 0 , k 1 ) , ( x 1 , k 1 ) , ( x 1 , k 2 ) , ( x 2 , k 2 ) , ( x 2 , k 3 ) , ( x 3 , k 3 ) ,..., ( x s 1 , k s ) , ( x s , k s ) are the e xact positions of the v ertices of a trajectory E s ( x ) in which all streets are straight and perpendicular and suppose that it v eries that x i < x j for all 0 i < j s . W ith this notation, the follo wing Proposition can be established. Pr oposition 1 Let P s = { x 0 = a, x 1 , ..., x s = b } be a partition of [ a, b ] and let E s ( x ) be a step function dened in [ a, b ] by (3). E s ( x ) = k 1 · χ [ x 0 ,x 1 ] + s X i =2 k i · χ ( x i 1 ,x i ] (3) with k i real numbers and χ [ c,d ] ( x ) the function dened by χ [ c,d ] ( x ) = 1 if x [ c, d ] and χ [ c,d ] ( x ) = 0 if x / [ c, d ] . F or all n 2 , if 3( b a ) n K δ ( s ) , being δ ( s ) = min 1 i s | x i x i 1 | and K 2 a posi ti v e inte ger , it follo ws that 1. | C n ( x ) E s ( x ) | 2 K ( M s m s ) n n if x [ a, b ] S s 1 i =1 x i δ ( s ) 3 , x i + δ ( s ) 3 2. | C n ( x ) [ k i (1 α x ) + k i +1 α x ] | 2 K ( M s m s ) n n if i = 1 , ..., s 1 and x x i δ ( s ) 3 , x i + δ ( s ) 3 where M s and m s are the maximum and the minimum of the k i , α x (0 , 1) is a real number which depends only on x and ( C n ) n is the sequence of RAFU functions associated to E s according with (1) and dened as (4). C n ( x ) = k 1 + s X i =2 [ k i k i 1 ] · F n, 2 ( x i 1 , x ) (4) being F n, 2 ( x i 1 , x ) for each i = 2 , ..., s the functions dened in (2) for p = 2 . A proof of Proposition 1 can be seen in [17]. The e xpression E r r or ( n ) = 2 K ( M s m s ) n n gi v e us the maxim um distance between the T axi Function C n ( x ) and the real trajectory E s ( x ) along the road centerline. In this sense, x ed a certain K , for an y n 2 we can kno w E r r or ( n ) and reciprocally . Int J Rob & Autom, V ol. 10, No. 4, December 2021 : 319 325 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Rob & Autom ISSN: 2089-4856 321 In practice we w ould w ork in this w ay: rst consider a maximum error bound E r r or ( n ) , then we calculate δ ( s ) and K = I n t [ l og 2 3( b a ) δ ( s ) ] and nally we nd the v al ue of n in order to determine the T axi function C n ( x ) that the robot must follo w . Example 1 Suppose that the trajectory that a car must follo w consists of 13 straight and perpendicular streets of 2275 m long in total and it is gi v en by means the step function E 7 ( x ) = 40 if 20 x 150 350 if 150 < x 600 200 if 600 < x 750 10 if 750 < x 925 80 if 925 < x 1100 250 if 1100 < x 1200 75 if 1200 < x 1400 In this case δ ( s ) = 100 and 3 ( b a ) n K 3 ( b a ) 2 K 100 = δ ( s ) holds for all n 2 when K = 6 According to Proposition 1 , for all x [20 , 1400] , | C n ( x ) E 7 ( x ) | E r r or ( n ) = 2 K ( M s m s ) n n = 2 6 (350 10) n n In Figure 1 we represent the e xact trajectory E 7 ( x ) (red color) and its continuous approximations C 40 ( x ) respecti v ely (blue color). 200 400 600 800 1000 1200 1400 50 100 150 200 250 300 350 Figure 1. Approximation to a step function In section 2 we ha v e applied the result to the particular case in which the journe y is dened by a step function, b ut in [10] we studied ho w t he RAFU functions can be used to approach an arbitrary discontinuous function. 3. TRAJECT OR Y DEFINED BY A CONTINUOUS FUNCTION Suppose that inside a machine we kno w the e xact positions of some points ( x 0 , f ( x 0 )) , ( x 1 , f ( x 1 )) , ..., ( x n 1 , f ( x n 1 )) and ( x n , f ( x n )) of a continuous trajectory f ( x ) . Denition 2 Let f be a function dened in [ a, b ] . The modulus of continuity of f , ω ( f , h ) , is the maximum of | f ( x ) f ( y ) | for all a x, y b , | x y | h . 3.1. Case of a unif orm net Let [ a, b ] be an interv al and suppose the case in which the v alues x i v erify x i = a + i · b a n for each i = 0 , ..., n . The taxi function (Alicia C. S ´ anc hez) Evaluation Warning : The document was created with Spire.PDF for Python.
322 ISSN: 2089-4856 Pr oposition 2 Let P n = { x 0 = a, x 1 , ..., x n = b } be a partition of [ a, b ] with x j = a + j · b a n , j = 0 , 1 ,..., n and let E n be the step function dened by (5). E n ( x ) = k 1 x [ a, x 1 ] k 2 x ( x 1 , x 2 ] ... k n x ( x n 1 , b ] k j R , j = 1 , ..., n (5) Let C n be the RAFU function ass ociated to E n dened as (4) by C n ( x ) = k 1 + P n j =2 [ k j k j 1 ] · F n, 2 ( x j 1 , x ) . Then, for all n 2 it follo ws that: 1. | C n ( x ) E n ( x ) | 2( M n m n ) n n if x [ a, b ] \ n 1 k =1 x k b a 3 n , x k + b a 3 n 2. | C n ( x ) [ k j (1 α x ) + k j +1 α x ] | 2( M n m n ) n n if x x j b a 3 n , x j + b a 3 n and j = 1 ,..., n 1 being M n and m n the maximum and the minimum of the k j respecti v ely and α x (0 , 1) a number that depends upon x . Theor em 1 Let f be a continuous function dened in [ a, b ] and let P n = { x 0 = a, x 1 , ..., x n = b } be a partition of [ a, b ] where x i = a + i · h for each i = 0 , ..., n and h = b a n . Then there e xists a sequence of radical functions ( C n ) n dened in [ a, b ] as (1) such that | C n ( x ) f ( x ) | 2 ( M m ) n n + ω ( f , h ) for all n 2 being M and m the maximum and the minimum of f in [ a, b ] respecti v ely and ω ( f , h ) its modulus of continuity . The proofs of Proposition 2 and Theorem 2 can be seen in [17]. According to these proofs, we dene the functions E n ( x ) as in Proposition 1 b ut taking into account that k 1 = f ( x 0 ) = f ( x 1 ) and k i = f ( x i ) for all i = 2 , ..., n and then its corresponding T axi functions C n ( x ) as in (4). In practice we w ould w ork in the same w ay that we ha v e mentioned in section 2 . Example 2 Gi v en a route dened by the continuous function f ( x ) f ( x ) = 75 if 20 x 80 ( x 200) 2 288 + 25 if 80 < x 200 5 x 850 6 if 200 < x 350 150 if 350 < x 500 ( x 500) 2 750 + 150 if 500 < x 800 Under the h ypothesis of Theorem 1 , we kno w that 2 ( M m ) n n + ω ( f , h ) 2 (150 25) n n + 5 6 · 800 20 n 952 n n In Figure 2 we sho w the results for n = 100 . 200 400 600 800 25 50 75 100 125 150 175 Figure 2. Approximation to a continuous function Int J Rob & Autom, V ol. 10, No. 4, December 2021 : 319 325 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Rob & Autom ISSN: 2089-4856 323 It is important to observ e that it is not easy to nd the error E r r or ( n ) in Theorem 1 because of he dif culty to obtain the term ω ( f , h ) . So, what can we do? In practice, we can calculate ∆( f ) = max 1 i n | f ( x i +1 ) f ( x i ) | and we could tak e this v alue instead of ω ( f , h ) . Some adv antages of the use of the T axi functions: The v alue of n for the functions E n ( x ) and C n ( x ) is the same. The stability of the T axi functions impro v es t he instability of other f amilies of approximating functions and this is an important contib ution of this w ork. This can be inferred from Proposition 2 where we can deduce that between each tw o consecuti v e v alues x r and x r +1 the T axi functions C n ( x ) tak e v alues between k r and k r +1 . The functions F n,p in (2) do not depend on the points x i b ut only on the subindices i (see p. 114 [12]). So, this reduces the calculations in v olv ed. The T axi functions C n ( x ) can be obtained with the only condition that the trajectory to approach is a continuous function. According to Proposition 2 , the m aximum error bound made with the T axi functions depends only on parameters that are kno wn. 3.2. Case of a non unif orm net Suppose the case in which the v alues x i form a non uniform net. In this case an analogous result to the T eorem 3 . 1 can be obtained. Theor em 2 Let P n = { x 0 = a, x 1 , ..., x s n = b } be a partition of [ a, b ] with δ ( s n ) = min 1 i s n | x i x i 1 | and ( s n ) = max 1 i s n | x i x i 1 | such that for all n 2 , 3( b a ) n K δ ( s ) ( s n ) h being h = b a n and K 2 a positi v e inte ger . Let f be a continuous function dened in [ a, b ] . Then, there e xists a sequence ( C n ) n dened in [ a, b ] as (1) for p = 2 such that | C n ( x ) f ( x ) | 2 K ( M m ) n n + ω ( f , h ) for all n 2 and x [ a, b ] being M and m the maximum and the minimum of f respecti v ely and ω ( f , h ) its modulus of continuity . A proof of Theorem 2 can be seen in [17]. T o obtain the T axi function C n ( x ) we w oul w ork as in the pre vious subsection. The same adv antages hold in this case too. 4. CASE OF AN ARBITRAR Y TRAJECT OR Y ON THE PLANE Suppose the information we kno w inside a machine is the e xact position of some v alues ( x 0 , T ( x 0 )) , ( x 1 , T ( x 1 )) , ..., ( x s 1 , T ( x s 1 )) and ( x s , T ( x s )) that belong to a continuous trajectory on the plane ( x, T ( x )) . Here we study what we can do when this trajectory is not a function. As ( x, T ( x )) is not a function, without loss of generality we can suppose that x 0 < ... < x i x i +1 for a certain i . There are tw o possible cases. 1). Case 1 . It v eries that x 0 < ... < x i x i +1 and T ( x i +1 ) < T ( x i ) . Then we use the equations of the rotation around the center ( x i , T ( x i )) and through the angle α = 90 º to turn the points ( x i +1 , T ( x i +1 )) , ..., ( x s , T ( x s )) into ( x i +1 , 2 , T ( x i +1 , 2 )) , ..., ( x s, 2 , T ( x s, 2 )) respecti v ely 1 x j , 2 T 2 ( x j , 2 ) = 1 0 0 x i + T ( x i ) 0 1 x i + T ( x i ) 1 0 · 1 x j T ( x j ) for all i + 1 j s . 2). Case 2 . It v eries that x 0 < ... < x i x i +1 and T ( x i +1 ) > T ( x i ) . Then we use the equations of the ro- tation around the center ( x i , T ( x i )) and through the angle α = 90 º to turn the points ( x i +1 , T ( x i +1 )) , ..., ( x s , T ( x s )) into ( x i +1 , 2 , T ( x i +1 , 2 )) , ..., ( x s, 2 , T ( x s, 2 )) respecti v ely The taxi function (Alicia C. S ´ anc hez) Evaluation Warning : The document was created with Spire.PDF for Python.
324 ISSN: 2089-4856 1 x j , 2 T 2 ( x j , 2 ) = 1 0 0 x i T ( x i ) 0 1 x i + T ( x i ) 1 0 · 1 x j T ( x j ) for all i + 1 j s . In this w ay , we will reproduce the cases 1 and 2 as man y times as necessary until the initial points ( x 0 , T ( x 0 )) , ( x 1 , T ( x 1 )) , ..., ( x s , T ( x s )) ha v e been turned into other points ( x 0 , T ( x 0 )) , ..., ( x s,q , T q ( x s,q )) in which x 0 < ... < x s,q . So, the initial trajectory ( x, T ( x )) becomes a continuous function F ( x ) to which we can apply the e xplained in section 3 . After then, we obtain the T axi function C n ( x ) associated to F ( x ) for a certain n and then we recalculate for each x [ a, b ] its the corresponding point of the real trajectory according to all the in v erse rotations that this x has tested in order to gi v e to the robot the real information to follo w . 5. CONCLUSION The RAFU method on approximation is an original approximation procedure. In this w ork we ha v e sho wed ho w this method can be used to solv e the lane k eeping and the lateral control problems in intelligent transport systems, specically in robot na vig ation. Gi v en a trajectory on the plane, the T axi functions can be useful to pro vide the necessary information to a v ehicle in order to go from a place to another one autonomously . The proposed method holds for an y trajectory and it does not depend on its re gularity . Moreo v er the error bound that the T axi function pro vides does not depend on an y unkno wn parameter or of the re gularity of the trajectory . Gi v en the conciseness of the results of this paper , we belie v e that these a v enues of research deserv e some attention. REFERENCES [1] N. M. Sarif, R. Ng adengon, H. A. Kadir , M. H. A. Jalil and K. Abidi, ”A discrete-time terminal sliding mode controller design for an autonomous underw ater v ehicle, IAES International Journal of Robotics and Automation (IJRA) , v ol. 10, no. 2, pp. 104-113, 2021, doi: 10.11591/ijra.v10i2.pp104-113. [2] W . K. Al-Azza wi, ”W ireless stepper m otor control and optimization based on rob ust control theory , IAES International Journal of Robotics and Automation (IJRA) , v ol. 10, no.2, pp. 144-148, 2021, doi: 10.11591/ijra.v10i2.pp144-148. [3] C, T . Nnodim, M. O. Aro w olo, B. D. Agboola, R. O. Ogundokun and M. K. Abiodun, ”Future trends in mechatronics, IAES International Journal of Robotics and Automation (IJRA) , v ol. 10, no.1, pp. 24-31, 2021, doi: 10.11591/ijra.v10i1.pp24-31. [4] H. Ab Ghani, et al ., ”Adv ances in lane marking detection algorithms for all-weather conditions, Interna- tional Journal of Electrical and Computer Engineering (IJECE) , v ol.11, no.4, pp. 3365-3373, 2021, doi: 10.11591/ijece.v11i4.pp3365-3373. [5] A. Khodayari, A. Ghaf f ari, S. Ameli and J. Flahatg ar , ”A historical re vie w on lateral and longitudinal control of autonomous v ehicle motions, 2010 International Conference on Mechanical and Electrical T echnology , 2010, pp. 421-429, doi: 10.1109/ICMET .2010.5598396. [6] S. G. Fernandez, et al ., ”Unmanned and autonomous ground v ehicle, International Journal of Electrical and Computer Engineering (IJECE) , v ol. 9, no. 5, pp. 4466-4472, 2019, doi: 10.11591/ijece.v9i5.pp4466- 4472. [7] Y . Lu and L. Bi, ”Combined Lateral and Longitudinal Control of EEG Signals-Based Brain-Controlled V ehicles, in IEEE T ransactions on Neural Systems and Rehabilitation Engineering , v ol. 27, no. 9, pp. 1732-1742, Sept. 2019, doi: 10.1109/TNSRE.2019.2931360. [8] M. S. Netto, S. Chaib and S. Mammar , ”Lateral adapti v e control for v ehicle lane k eeping, Proceedings of the 2004 American Control Conference , 2004, pp. 2693-2698 v ol.3, doi: 10.23919/A CC.2004.1383872. [9] Z. Gyenes and E. G. Sz ´ adeczk y-Kardoss, ”A no v el concept of lane-k eeping algorithms for mobile robots, 2019 IEEE 17th International Symposium on Intelligent Systems and Informatics (SISY) , 2019, pp. 000047-000052, doi: 10.1109/SISY47553.2019.9111614. [10] A. C. S ´ anchez, ”The narro w border between a discontinuous function and a continuous func- tion, SCIREA Journal of mathematics , v ol. 5, no. 3, pp. 32-43. 2020. [Online]. A v ailable: http://www .scirea.or g/journal/P aperInformation?P aperID=3754. Int J Rob & Autom, V ol. 10, No. 4, December 2021 : 319 325 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Rob & Autom ISSN: 2089-4856 325 [11] A. C. S ´ anchez, ”The RAFU remainder in T aulor’ s formula, The Australian Journal of Mathemati- cal Analysis and Applications , v ol. 18, no 2, p. 7, 2021. [Online]. A v ailable: https://ajmaa.or g/cgi- bin/paper .pl?string=v18n2/V18I2P7.te x. [12] E. Corbacho, ”Uniform approximation with radical functions, SeMA Journal , v ol . 58, no. 1, pp. 97-122, 2012, doi: 10.1007/BF03322606. [13] E. Corbacho, ”A RAFU linear space uniformly dense in C [a, b], Applied General T opol o gy , v ol. 14, no. 1, pp. 53-60, 2013, doi: 10.4995/agt.2013.1617. [14] E. Corbacho, Introduction to the RAFU met ho d on approximation , LAP Lambert Academic Publishing, 2015. [15] E. Corbacho, ”Approximation in dif ferent smoothness s paces with the RAFU method”, Applied General T opology , v ol. 15, no. 2, pp. 221-228, 2014, doi: 10.4995/agt.2014.2139. [16] E. Corbacho, ”Simultaneous approximation with the RAFU method”, Journal of Mathematical Inequali- ties , v ol. 10, no. 1, pp. 219-231, 2016, doi: 10.7153/jmi-10-19. [17] E. Corbacho, ”Uniform reconstruction of continuous functions with the RAFU method”, Applied General T opology , v ol. 18, no. 2, pp. 361-375, 2017, doi: 10.4995/agt.2017.7263. [18] S. Arianto, H. W ibo w o, W . Suharso, R. No vidianto and Dani Harmanto, ”3D Reconstruction using con v olution smooth method, Bulletin of Electrical Engineering and Informatics , v ol. 10, no. 3, pp. 1337- 1344, 2021, doi: 10.11591/eei.v10i3.1991. [19] P . Me g antoro, B. A. Pramudita, P . V igneshw aran, A. Y urianta and Hendra Ari W inarno, ”Real- time monitoring system for weather and air pollutant measurement ith HTML-based UI applica- tion, Bulletin of Electrical Engineering and Informatics , v ol. 10, no. 3, pp. 1669–1677, 2021, doi: 10.11591/eei.v10i3.3030. [20] A. V . V itianingsih, N. Suryana and Z. Othman, ”Spatial analysis model for traf c accident-prone roads classication: a proposed frame w ork, IAES International Journal of Articial Intelligence (IJ-AI) , v ol. 10, no. 2, pp. 365-373, 2021, doi: 10.11591/ijai.v10.i2.pp365-373. [21] N. T . Morallo, ”V ehicle track er system design based on GSM and GPS interf ace using arduino as plat- form, Indonesian Journal of Electrical Engineering and Computer Science , v ol. 23, no. 1, pp. 258–264, 2021, doi: 10.11591/ijeecs.v23.i1.pp258-264. [22] E. A. Lisang an and S. C. Sumarta, ”An impro v ed fuzzy ant colon y system for route selection based on real time traf c condition, IAES International Journal of Articial Intelligence (IJ-AI) , v ol. 10, no. 1, pp. 51-59, 2021, doi: 10.11591/ijai.v10.i1.pp51-59. The taxi function (Alicia C. S ´ anc hez) Evaluation Warning : The document was created with Spire.PDF for Python.