Foot/Ankle Prostheses Design Approach Based on Scientometric and Patentometric Analyses

: There are different alternatives when selecting removable prostheses for below the knee amputated patients. The designs of these prostheses vary according to their different functions. These prostheses designs can be classiﬁed into Energy Storing and Return (ESAR), Controlled Energy Storing and Return (CESR), active, and hybrid. This paper aims to identify the state of the art related to the design of these prostheses of which ESAR prostheses are grouped into ﬁve types, and active and CESR are categorized into four groups. Regarding patent analysis, 324 were analyzed over the last six years. For scientiﬁc communications, a bibliometric analysis was performed using 104 scientiﬁc reports from the Web of Science in the same period. The results show a tendency of ESAR prostheses designs for patents (68%) and active prostheses designs for scientiﬁc documentation (40%).


Introduction
Below-knee amputation (BKA) is a surgical procedure that mainly originates from trauma, diabetes, and peripheral vascular diseases [1]. While it is estimated that an average person walks about 6500 steps per day, current trends suggest that 10,000 steps per day represent a healthy lifestyle [2] for which a suitable prosthesis is necessary for a BKA patient in order to achieve a complete user reintegration to his/her pre-amputation activities. These designs should adapt to different patient's activities.
In scientific documents, there is wide confusion with the terms prosthesis, prosthetic, and prostheses; prosthetic is the process to manufacture an artificial member (AM), prosthesis a component of the AM, and prostheses are all the components that make up an AM. From patents and scientific document searches, the term prosthesis is more commonly used; in this paper, prostheses and prosthesis will be used interchangeably.
Understanding the functioning of these prostheses is necessary to identify the foot movements: internal-external axial rotation, eversion-inversion, dorsiflexion (DF), and plantarflexion (PF), as shown in Figure 1. The forces acting on the human foot are distributed with 60% towards the heel and 40% towards the phalanges. The loads are distributed between the heel and the metatarsals to the fourth and fifth phalanges and towards the big toe to the second and third phalanges [3].
In order to improve and develop ankle/foot prostheses, it is necessary to know and understand present-day solutions to walking and running for BKA patients (and the people behind those solutions), so our designs meet both user and technical requirements. A state-of-the-art analysis of BKA prostheses is performed in this research. Foot prostheses can be classified as follows: • Ankle-cushion heel (SACH-foot): This was developed in the 1950s and incorporated a compressible heel that dampens the impact on the ground while emulating a plantarflexion movement. This type of prosthesis is used for its relatively low cost and weight [4]. • ESAR, also known as ESR, was developed in the 1980s. This type of prosthesis uses a foot-modeled plate (usually carbon fiber made) that stores elastic potential energy and progressively releases it as kinetic energy [5]. • CESR prostheses aim to capture the energy that is dissipated during a gait impact.
On the loading phase of stance, energy is stored by a spring and locked. Then, this energy is timely released during the terminal stance of walking using microelectronic components [5]. • Active prostheses are considered state-of-the-art prostheses due to the use of actuators, microcontrollers, or other electronic devices; usually, these work using ESAR foot systems combined with some external elements such as actuators or other electronic components. These prostheses have better control and stability during a walk cycle [6].
In the next section, it is explained how the investigation was performed for both patents and scientific communications. The Result Section presents a discussion about a new prosthesis classification according to this investigation, main authors, countries, and keywords analyzed. In the discussion Section, findings and other designs of prosthesis designs are disclosed.

Search Method
BKA prostheses designs vary in form and functions, so in order to understand the way these designs work, extensive patents and scientific documentation searches were performed.

Used Keywords
For the patents and scientific communications searches, the following boolean operations were used under the International Patent Classification (IPC) A61F2 belonging to artificial substitutes or replacements for parts of the body: ((Ankle OR foot) AND (prosthetic OR prosthesis OR artificial)). Dates ranges were set from 2014 to 2020. For the patent analysis, 9526 documents were found. A scientific communications search provided 406 results. Figure 2 shows the results filtered on different search engines and the total number of documents obtained in every stage, among which The Lens was the most effective. Foot prostheses can be classified as follows: • Ankle-cushion heel (SACH-foot): This was developed in the 1950s and incorporated a compressible heel that dampens the impact on the ground while emulating a plantarflexion movement. This type of prosthesis is used for its relatively low cost and weight [4]. • ESAR, also known as ESR, was developed in the 1980s. This type of prosthesis uses a foot-modeled plate (usually carbon fiber made) that stores elastic potential energy and progressively releases it as kinetic energy [5]. • CESR prostheses aim to capture the energy that is dissipated during a gait impact.
On the loading phase of stance, energy is stored by a spring and locked. Then, this energy is timely released during the terminal stance of walking using microelectronic components [5]. • Active prostheses are considered state-of-the-art prostheses due to the use of actuators, microcontrollers, or other electronic devices; usually, these work using ESAR foot systems combined with some external elements such as actuators or other electronic components. These prostheses have better control and stability during a walk cycle [6].
In the next section, it is explained how the investigation was performed for both patents and scientific communications. The Result Section presents a discussion about a new prosthesis classification according to this investigation, main authors, countries, and keywords analyzed. In the discussion Section, findings and other designs of prosthesis designs are disclosed.

Search Method
BKA prostheses designs vary in form and functions, so in order to understand the way these designs work, extensive patents and scientific documentation searches were performed.

Used Keywords
For the patents and scientific communications searches, the following boolean operations were used under the International Patent Classification (IPC) A61F2 belonging to artificial substitutes or replacements for parts of the body: ((Ankle OR foot) AND (prosthetic OR prosthesis OR artificial)). Dates ranges were set from 2014 to 2020. For the patent analysis, 9526 documents were found. A scientific communications search provided 406 results. Figure 2 shows the results filtered on different search engines and the total number of documents obtained in every stage, among which The Lens was the most effective.

Patent Search
For the patent search, five different search engines were used, of which four were free-source, and one was paid. The databases were Derwent analytics (842 results), Espacenet (86 results), Google patents (5539 results), Patentscope (2281 results), and The Lens (778 results), with a total of 9526 results (see Figure 2). An initial filter was applied directly to the search engines where undesired categories and keywords were removed, in addition to a manual selection of patents directly on the website.

Patent Search
For the patent search, five different search engines were used, of which four were free-source, and one was paid. The databases were Derwent analytics (842 results), Espacenet (86 results), Google patents (5539 results), Patentscope (2281 results), and The Lens (778 results), with a total of 9526 results (see Figure 2).
An initial filter was applied directly to the search engines where undesired categories and keywords were removed, in addition to a manual selection of patents directly on the website.
Subsequently, data cleaning was performed using Open refine ® . The second filter was applied to eliminate duplicates, IPC categories that did not correspond, and keywords such as heart, valve, elbow, Arthroplasty, and Orthosis. An individual selection of the patents was made, and the unwanted results were eliminated. The remaining patents were as follows: Patentscope (369), Google patents (390), Espacenet (55), Derwent analytics (546), and The Lens (309), resulting in 1669 patents.
Based on a third filter, the results of all databases were merged, and keywords such as knee, orthosis, and tibia were eliminated. Duplicated results were filtered, and the remaining patents were individually analyzed for a total result of Derwent analytics (70), Espacenet (12), Google patents (19), Patentscope (72), and The Lens (151), resulting in 324 patents directly related to ankle and foot prostheses. From Figure 2, it can be observed that although Google patents and Patentscope were the ones with more results, these contained a higher number of duplicates or undesired data. Subsequently, data cleaning was performed using Open refine ® . The second filter was applied to eliminate duplicates, IPC categories that did not correspond, and keywords such as heart, valve, elbow, Arthroplasty, and Orthosis. An individual selection of the patents was made, and the unwanted results were eliminated. The remaining patents were as follows: Patentscope (369), Google patents (390), Espacenet (55), Derwent analytics (546), and The Lens (309), resulting in 1669 patents.
Based on a third filter, the results of all databases were merged, and keywords such as knee, orthosis, and tibia were eliminated. Duplicated results were filtered, and the remaining patents were individually analyzed for a total result of Derwent analytics (70), Espacenet (12), Google patents (19), Patentscope (72), and The Lens (151), resulting in 324 patents directly related to ankle and foot prostheses. From Figure 2, it can be observed that although Google patents and Patentscope were the ones with more results, these contained a higher number of duplicates or undesired data.

Scientific Communications Search
For the literature analysis, the same keywords as for the patents' search were applied in the Web of Science (WOS), obtaining 406 documents related to foot/ankle prostheses. The first filter was performed directly on the website, removing undesired keywords for a total of 136 documents. Subsequently, a second filter was applied, deleting repeated and undesired results. An individual document selection was made, resulting in 97 results. Finally, a bibliometric analysis was performed using data recovery software (R studio ® ) and a complement for bibliometric analysis (Bibliometrix ® ).

Patentometric Analysis
Among the 324 results obtained, 208 results match prostheses designs, 51 match prosthetic mechanisms (motion blocking systems, aids to align prostheses, etc.), 22 match sockets, 11 match aesthetic covers, and 10 match joints. In total, 22 results are associated with methodologies (manufacturing methods, design methods, tests). Figure 3 shows these results; the number of prosthesis designs suggests a high interest in the development of new solutions for BKA amputees.

Scientific Communications Search
For the literature analysis, the same keywords as for the patents' search were applied in the Web of Science (WOS), obtaining 406 documents related to foot/ankle prostheses. The first filter was performed directly on the website, removing undesired keywords for a total of 136 documents. Subsequently, a second filter was applied, deleting repeated and undesired results. An individual document selection was made, resulting in 97 results. Finally, a bibliometric analysis was performed using data recovery software (R studio ® ) and a complement for bibliometric analysis (Bibliometrix ® ).

Patentometric Analysis
Among the 324 results obtained, 208 results match prostheses designs, 51 match prosthetic mechanisms (motion blocking systems, aids to align prostheses, etc.), 22 match sockets, 11 match aesthetic covers, and 10 match joints. In total, 22 results are associated with methodologies (manufacturing methods, design methods, tests). Figure 3 shows these results; the number of prosthesis designs suggests a high interest in the development of new solutions for BKA amputees. Among the results, 95 refer to foot prostheses, 65 to ankle prostheses, and 48 to a combination of both, of which 182 are removable, and 26 are osseointegrated. In this investigation, only removable prostheses are considered. Table 1 shows the selected patents, the technology used, and the type of prostheses. Among removable prostheses, 135 are mechanical or propelled with the body, hydraulic (18 results), and electronic or active (29 results). These results are distributed among ESAR, CESR, active, and hybrid (which did not match any of the aforementioned technologies or they are a combination of two or more categories). From Figure 4, it can be observed that for electronic prostheses, 17 are active, three are CERS (use a controlled energy return without the use of complex devices), one is ESAR, and eight are hybrid. For hydraulic prostheses, four use electronic components, three are based on CERS, three on ESAR, and eight are a combination of three or more categories. For mechanical prostheses, 94 use ESAR systems exclusively, 26 combine different technologies (but mostly are mechanical), 13 are CERS (energy return is controlled using only mechanical devices), and two use actuators to release the energy. Among the results, 95 refer to foot prostheses, 65 to ankle prostheses, and 48 to a combination of both, of which 182 are removable, and 26 are osseointegrated. In this investigation, only removable prostheses are considered. Table 1 shows the selected patents, the technology used, and the type of prostheses. Among removable prostheses, 135 are mechanical or propelled with the body, hydraulic (18 results), and electronic or active (29 results). These results are distributed among ESAR, CESR, active, and hybrid (which did not match any of the aforementioned technologies or they are a combination of two or more categories). From Figure 4, it can be observed that for electronic prostheses, 17 are active, three are CERS (use a controlled energy return without the use of complex devices), one is ESAR, and eight are hybrid. For hydraulic prostheses, four use electronic components, three are based on CERS, three on ESAR, and eight are a combination of three or more categories. For mechanical prostheses, 94 use ESAR systems exclusively, 26 combine different technologies (but mostly are mechanical), 13 are CERS (energy return is controlled using only mechanical devices), and two use actuators to release the energy.
Applicants and inventors in the databases were considered. Otto Bock Health Co. and Clausen Arinbjorn V. are the main applicants with ten and eight patents, respectively, from 2014 to 2020. Figure 5 shows the main applicants for BKA prostheses.  Applicants and inventors in the databases were considered. Otto Bock Health Co. and Clausen Arinbjorn V. are the main applicants with ten and eight patents, respectively, from 2014 to 2020. Figure 5 shows the main applicants for BKA prostheses.

Scientometric Analysis
After the final filter was applied, 98 scientific documents directly related to ankle/foot prostheses were selected; results are shown in Table 2. Keywords were analyzed resulting in the top 10: gait (frequency = 15 articles), prosthesis (frequency =14 articles), prosthetics (frequency =13 articles), amputation (frequency =11 articles), biomechanics (frequency =11 articles), ankle (frequency = 8 articles), transtibial (frequency = 8 articles) prosthetic foot (frequency = 7 articles), powered prosthesis (frequency = 6 articles), and gait analysis (frequency = 5 articles). This means there is a major trend in developing prostheses devices compared with gait studies or the creation of new methodologies.
From the information obtained by the scientific documents, several aspects must be considered when designing a new prosthesis, such as aesthetics, which allows empathy between the users and their prosthesis [1], a size that permits the use of footwear, a mass corresponding to 2.5% of bodyweight [160] (literature shows an average of 2.5 kg for a 75 kg person), an ankle torque corresponding to 100-140 Nm, an ankle power between 250-300 W, and a device capable of storing and releasing energy (5-9 J) On the authors' part, Lefeber D. and Vanderborght B. are the top authors (11 articles each). Nevertheless, Hugh M. Herr is the most cited author in this field, with five of the most cited articles. Applicants and inventors in the databases were considered. Otto Bock Health Co. and Clausen Arinbjorn V. are the main applicants with ten and eight patents, respectively, from 2014 to 2020. Figure 5 shows the main applicants for BKA prostheses.    [190] Sun, Jinming Clinical Study 2014 [191] Wezenberg, Daphne Comparative Study 2014 [192] Nickel, Eric Component Design 2014 [193] Mulder, Inge A. Foot Prosthesis Design 2014 [194,195] Safaeepour, Zahra Powered Ankle/foot Prosthesis design 2014 [196] Zhu, Jinying Powered Ankle/foot Prosthesis design 2014 [197,198] Ko, Chang-Yong Clinical Study 2014-2016 [199,200] Cherelle   Table 3 shows, in order, the most cited articles, and Figure 6 shows the most relevant authors in scientific documentation. On the authors' part, Lefeber D. and Vanderborght B. are the top authors (11 articles each). Nevertheless, Hugh M. Herr is the most cited author in this field, with five of the most cited articles. Table 3 shows, in order, the most cited articles, and Figure 6 shows the most relevant authors in scientific documentation.  The United States (US) is the most productive country (46 documents), followed by Belgium (seven documents) and China (five documents). Some documents showed multiple country collaborations (Figure 7). There is a clear relation between authors, journals, and countries. For example, most of the documents submitted in the US are from IEEE magazines and Plos One; meanwhile, Europe tends to apply to Prosthetic and Orthotic international and the American society of mechanical engineers (ASME).
The United States (US) is the most productive country (46 documents), followed by Belgium (seven documents) and China (five documents). Some documents showed multiple country collaborations (Figure 7). There is a clear relation between authors, journals, and countries. For example, most of the documents submitted in the US are from IEEE magazines and Plos One; meanwhile, Europe tends to apply to Prosthetic and Orthotic international and the American society of mechanical engineers (ASME).

Device Classification
From the selected patents and scientific documentation, a new ankle/foot prosthesis classification has been created besides ESAR, CERS, and active, based on its components and prosthesis functions.
ESAR prostheses are categorized into five different designs (see Figure 8). CERS and active categories are merged and divided into five different categories. There are some unique designs whose components cannot be grouped; these will be discussed individually.

Device Classification
From the selected patents and scientific documentation, a new ankle/foot prosthesis classification has been created besides ESAR, CERS, and active, based on its components and prosthesis functions.
ESAR prostheses are categorized into five different designs (see Figure 8). CERS and active categories are merged and divided into five different categories. There are some unique designs whose components cannot be grouped; these will be discussed individually. From the previous analyses, it can be determined that the general form for ESAR prosthesis is similar to the one illustrated in Figure 8A and mostly differs in form; sometimes, a single talon plate is aggregated, or the disposition of the plates may vary. In other cases, as in Figure 8B, the center of mass is moved, and the plates are rearranged. In the variation represented by Figure 8C, the foot plates are divided, so the prosthesis emulates eversion and inversion movements. In Figure 8D, some polymeric cushions are aggregated, replacing the use of extra plates. Figure 8E shows the usage of different types of damping systems (springs, actuators, etc.) that replace some plates. All of these designs use pyramid adapters as a connection between the prosthesis and transtibial components. There are some variations for ESAR prostheses that use a simple plate arrangement to adjust the return of energy (see Figure 9A). Other designs use a single spring bar that regulates the energy storage/release (see Figure 9B). For CERS prosthesis, the model by Endo Ken [129] (see Figure 10) considers a locking mechanism that preserves the energy storage in the spring. This energy is released upon the foot movement during the terminal stance. This impulse, in combination with the ESAR foot, provides necessary torque during the walk cycle. From the previous analyses, it can be determined that the general form for ESAR prosthesis is similar to the one illustrated in Figure 8A and mostly differs in form; sometimes, a single talon plate is aggregated, or the disposition of the plates may vary. In other cases, as in Figure 8B, the center of mass is moved, and the plates are rearranged. In the variation represented by Figure 8C, the foot plates are divided, so the prosthesis emulates eversion and inversion movements. In Figure 8D, some polymeric cushions are aggregated, replacing the use of extra plates. Figure 8E shows the usage of different types of damping systems (springs, actuators, etc.) that replace some plates. All of these designs use pyramid adapters as a connection between the prosthesis and transtibial components.
There are some variations for ESAR prostheses that use a simple plate arrangement to adjust the return of energy (see Figure 9A). Other designs use a single spring bar that regulates the energy storage/release (see Figure 9B).
For CERS prosthesis, the model by Endo Ken [129] (see Figure 10) considers a locking mechanism that preserves the energy storage in the spring. This energy is released upon the foot movement during the terminal stance. This impulse, in combination with the ESAR foot, provides necessary torque during the walk cycle. There are some variations for ESAR prostheses that use a simple plate arrangement to adjust the return of energy (see Figure 9A). Other designs use a single spring bar that regulates the energy storage/release (see Figure 9B). For CERS prosthesis, the model by Endo Ken [129] (see Figure 10) considers a locking mechanism that preserves the energy storage in the spring. This energy is released upon the foot movement during the terminal stance. This impulse, in combination with the ESAR foot, provides necessary torque during the walk cycle. Active prostheses can be categorized by the components they use into three types: Multi-Array Prostheses (MAP), Low Powered Prostheses (LPP), and Controlled Adaptative Stiffness (CAS). For MAP, the form is similar to the one shown in Figure 11. It uses an ESAR composite foot (E), and a DC motor (A), usually a 200 W Maxon ® connected to a ball-screw transmission (C) that moves the linkage system (D) upward/downward and  For CERS prosthesis, the model by Endo Ken [129] (see Figure 10) considers a mechanism that preserves the energy storage in the spring. This energy is releas the foot movement during the terminal stance. This impulse, in combination ESAR foot, provides necessary torque during the walk cycle. Active prostheses can be categorized by the components they use into thr Multi-Array Prostheses (MAP), Low Powered Prostheses (LPP), and Controlled tive Stiffness (CAS). For MAP, the form is similar to the one shown in Figure 11. I ESAR composite foot (E), and a DC motor (A), usually a 200 W Maxon ® conne ball-screw transmission (C) that moves the linkage system (D) upward/downw Active prostheses can be categorized by the components they use into three types: Multi-Array Prostheses (MAP), Low Powered Prostheses (LPP), and Controlled Adaptative Stiffness (CAS). For MAP, the form is similar to the one shown in Figure 11. It uses an ESAR composite foot (E), and a DC motor (A), usually a 200 W Maxon ® connected to a ball-screw transmission (C) that moves the linkage system (D) upward/downward and converts motor rotary motion into linear motion. In some cases, the motor is located instead of the spring (G) and connected to (C) using a timing belt. The linkage system (D) is in charge of connecting different mechanisms and allows plantarflexion and dorsiflexion movements; it may be composed of cables and/or pulleys, a bar mechanism, or crank sliders. F and G, depending on the prostheses, represent springs or actuators (pneumatic, electric, or hydraulic), for which torque varies from 100 to 140 Nm. Sometimes a parallel spring is aggregated due to the demanding torque requirements, and it aims to reduce the loads supported by the linkage system. Spring (G) saves energy during plantarflexion and dorsiflexion and supplements it during the swing phase. Housing (B) allocates all the electronic systems and provides stability to the system. The pyramid adapter (H) provides a connection between the transtibial components and the prosthesis. Some models have a lock mechanism, so the prosthesis could be used in a passive mode. See Figures 11-15 converts motor rotary motion into linear motion. In some cases, the motor is located instead of the spring (G) and connected to (C) using a timing belt. The linkage system (D) is in charge of connecting different mechanisms and allows plantarflexion and dorsiflexion movements; it may be composed of cables and/or pulleys, a bar mechanism, or crank sliders. F and G, depending on the prostheses, represent springs or actuators (pneumatic, electric, or hydraulic), for which torque varies from 100 to 140 Nm. Sometimes a parallel spring is aggregated due to the demanding torque requirements, and it aims to reduce the loads supported by the linkage system. Spring (G) saves energy during plantarflexion and dorsiflexion and supplements it during the swing phase. Housing (B) allocates all the electronic systems and provides stability to the system. The pyramid adapter (H) provides a connection between the transtibial components and the prosthesis. Some models have a lock mechanism, so the prosthesis could be used in a passive mode. See Figures 11-15. Figure 11. MAP active ankle-foot prosthesis.
Another powered prosthesis design is the LPP shown in Figure 12. It aims to reduce the necessary power required by the actuators. It contains different Footplates (G and C), which in some designs (similar to the AMP Foot 2.1 [199]) are merged into a single plate.

Other Designs
Some designs do not correspond to the categories previously described. These designs are the pneumatic foot prosthesis by Huang et al. [189] (see Figure 14A), where DF and PF are managed by two artificial muscles each, so stiffness and PF torque are easier to control. It is capable of emulating 3 DOF and is controlled via a desktop computer. Another design is the two DOF cable-driven ankle-foot prosthesis by Ficanha et al. [213], where instead of using pneumatic systems, it uses pulleys and Bowden cables that are externally controlled by two motors (Maxon EC-4), see Figure 14B. Both systems have an external power source and are capable of emulating foot eversion and inversion movements.

Other Designs
Some designs do not correspond to the categories previously described. These designs are the pneumatic foot prosthesis by Huang et al. [189] (see Figure 14A), where DF and PF are managed by two artificial muscles each, so stiffness and PF torque are easier to control. It is capable of emulating 3 DOF and is controlled via a desktop computer. Another design is the two DOF cable-driven ankle-foot prosthesis by Ficanha et al. [213], where instead of using pneumatic systems, it uses pulleys and Bowden cables that are externally controlled by two motors (Maxon EC-4), see Figure 14B. Both systems have an external power source and are capable of emulating foot eversion and inversion movements.

Other Designs
Some designs do not correspond to the categories previously described. These designs are the pneumatic foot prosthesis by Huang et al. [189] (see Figure 14A), where DF and PF are managed by two artificial muscles each, so stiffness and PF torque are easier to control. It is capable of emulating 3 DOF and is controlled via a desktop computer. Another design is the two DOF cable-driven ankle-foot prosthesis by Ficanha et al. [213], where instead of using pneumatic systems, it uses pulleys and Bowden cables that are externally controlled by two motors (Maxon EC-4), see Figure 14B. Both systems have an external power source and are capable of emulating foot eversion and inversion movements.  Another powered prosthesis design is the LPP shown in Figure 12. It aims to reduce the necessary power required by the actuators. It contains different Footplates (G and C), which in some designs (similar to the AMP Foot 2.1 [199]) are merged into a single plate. In another case such as the VSPA Foot [245], footplates (G) are individually controlled, allowing eversion-inversion movements; the DC motor (A) is located in a Housing (J) and rotates the Ball screw transmission (B), which moves the Footplate (C) up or down, allowing plantarflexion and dorsiflexion movement. Heel (D) may be composed of a flexible plate; ankle stiffness is provided by Springs (H) and (E). Depending on the model, two Springs (H) are used when there are individually controlled Footplates, and Spring (E) is used when (G) and (C) are merged. In this case, Spring (E) is attached directly to Footplate (C). Spring (E) is elongated using a Pulley system (F) connected to the Footplate (C). The pyramid adapter (I) provides a connection between the transtibial components and the prosthesis. Designs for this model use an external power supply that is not integrated into the main prosthesis body. Another case is the robotic foot prosthesis made by Lapre [229]. This device aims to actively align the foot during different stances of the gait cycle using a four-bar linkage system to rotate and translate the foot with the use of a single actuator. It works using an ESAR foot and a DC motor (Maxon ® EC- 30 200 W) that moves a Ball screw transmission via a belt drive. As this actuator system (motor and ball screw) contracts, it extends and shifts the foot center (see Figure 15).

ESAR Analysis
Most of the active prostheses use ESAR foot to generate enough power to initiate the gait cycle. From the patentometric and scientometric analysis, it is evident that types A, B, and C are the most used (see Figure 8). A structural analysis was performed to make a comparison between these types. Carbon-fiber footplates and a concrete floor were used. A load of 785 N was applied on the prosthesis upper faces obtaining a maximum deformation on the Y-axis of 0.63, 0.33, and 0.67 mm for types A, B, and C, respectively (see Figure 16). Meanwhile, deformations on A and C mostly occur on the ankle; B shows major flexibility along the foot. The red color shows maximum displacements on the foot connection with the body, but blue shows no deformation.
According to the structural analysis, B tends to offer major elastic energy compared to A and C, as shown in the instep colored in green/blue. To compare the effectiveness during a walk cycle on uneven terrain, prostheses A, B, and C were analyzed using the same velocity and loads. Figure 17 shows a clear advantage of (C) over the other two models, thanks to the uneven deformation on its divided footplates, as shown for the displacement colored in red. CAS prostheses (see Figure 13) are mainly based on an ESAR foot (D), and in some cases complemented with a Cushion (E). The main goal of this prosthesis is the modulation of the stiffness during different stages of a gait cycle. This is granted by moving a Slider (G) along the length of the foot. Depending on the gait cycle, this slider moves forward and backward, providing the necessary stiffness to adapt to different situations such as walking, running, or climbing stairs, and it is controlled by a DC motor (C). A linkage system could be provided by a Ball screw transmission (F) or pulleys and belts. Motor (C) could be programmed to adapt to different activities. Housing (B) provides support for all the components and allows one degree of freedom (DOF) for the foot. The pyramid adapter (A) provides a connection between the transtibial components and the prosthesis.

Other Designs
Some designs do not correspond to the categories previously described. These designs are the pneumatic foot prosthesis by Huang et al. [189] (see Figure 14A), where DF and PF are managed by two artificial muscles each, so stiffness and PF torque are easier to control. It is capable of emulating 3 DOF and is controlled via a desktop computer. Another design is the two DOF cable-driven ankle-foot prosthesis by Ficanha et al. [213], where instead of using pneumatic systems, it uses pulleys and Bowden cables that are externally controlled by two motors (Maxon EC-4), see Figure 14B. Both systems have an external power source and are capable of emulating foot eversion and inversion movements.
Another case is the robotic foot prosthesis made by Lapre [229]. This device aims to actively align the foot during different stances of the gait cycle using a four-bar linkage system to rotate and translate the foot with the use of a single actuator. It works using an ESAR foot and a DC motor (Maxon ® EC-30 200 W) that moves a Ball screw transmission via a belt drive. As this actuator system (motor and ball screw) contracts, it extends and shifts the foot center (see Figure 15).

ESAR Analysis
Most of the active prostheses use ESAR foot to generate enough power to initiate the gait cycle. From the patentometric and scientometric analysis, it is evident that types A, B, and C are the most used (see Figure 8). A structural analysis was performed to make a comparison between these types. Carbon-fiber footplates and a concrete floor were used. A load of 785 N was applied on the prosthesis upper faces obtaining a maximum deformation on the Y-axis of 0.63, 0.33, and 0.67 mm for types A, B, and C, respectively (see Figure 16). Meanwhile, deformations on A and C mostly occur on the ankle; B shows major flexibility along the foot. The red color shows maximum displacements on the foot connection with the body, but blue shows no deformation.

ESAR Analysis
Most of the active prostheses use ESAR foot to generate enough power to initiate the gait cycle. From the patentometric and scientometric analysis, it is evident that types A, B, and C are the most used (see Figure 8). A structural analysis was performed to make a comparison between these types. Carbon-fiber footplates and a concrete floor were used. A load of 785 N was applied on the prosthesis upper faces obtaining a maximum deformation on the Y-axis of 0.63, 0.33, and 0.67 mm for types A, B, and C, respectively (see Figure 16). Meanwhile, deformations on A and C mostly occur on the ankle; B shows major flexibility along the foot. The red color shows maximum displacements on the foot connection with the body, but blue shows no deformation.
According to the structural analysis, B tends to offer major elastic energy compared to A and C, as shown in the instep colored in green/blue. To compare the effectiveness during a walk cycle on uneven terrain, prostheses A, B, and C were analyzed using the same velocity and loads. Figure 17 shows a clear advantage of (C) over the other two models, thanks to the uneven deformation on its divided footplates, as shown for the displacement colored in red. According to the structural analysis, B tends to offer major elastic energy compared to A and C, as shown in the instep colored in green/blue.
To compare the effectiveness during a walk cycle on uneven terrain, prostheses A, B, and C were analyzed using the same velocity and loads. Figure 17 shows a clear advantage of (C) over the other two models, thanks to the uneven deformation on its divided footplates, as shown for the displacement colored in red.

Conclusions
The number of results per database does not reflect the effectiveness of each search engine. For this research, priority was given to search engines that provided useful data such as direct links to patents, the inventor's name, and IPC codes. Nevertheless, there are some difficulties with some of them, such as the lack of options for filtering results or IPC categories, among others. Besides, some applicants may be included in the name of their companies (for example, Herr Hugh in Massachusetts Institute of Technology); this is because some search engines only show the applicant/owner's name instead of the inventor. In some cases, there is a lack of consistency between the author's names in different patents (for example, Smith Keith and Smith, Keith, B.); these kinds of inconsistencies were clustered, but still, results could not be entirely precise.
The United States has 56% of patent applications and 34% of scientific documents registered. These results do not necessarily display that they produce most of the knowledge on this topic, but because of the language, most of the search engines are capable of accessing the data, unlike languages such as Spanish, Chinese, or languages spoken in India. Therefore, some designs could remain undiscovered for this investigation.
Based on the obtained results, it can be established that for this study, the effectiveness per search engine is as follows: Derwent 8.4%, Google patents 0.34%, Patentscope 3.2%, The Lens 19.9%, and Espacenet 13.95%.
The classification of the 208 prosthesis patents related to prostheses designs was obtained according to the main technology used; results show that the ESAR mechanical prosthesis is the main patent object by 44%, although claims are different for each one. All of them can be classified based on the five ESAR categories presented in this document.

Conclusions
The number of results per database does not reflect the effectiveness of each search engine. For this research, priority was given to search engines that provided useful data such as direct links to patents, the inventor's name, and IPC codes. Nevertheless, there are some difficulties with some of them, such as the lack of options for filtering results or IPC categories, among others. Besides, some applicants may be included in the name of their companies (for example, Herr Hugh in Massachusetts Institute of Technology); this is because some search engines only show the applicant/owner's name instead of the inventor. In some cases, there is a lack of consistency between the author's names in different patents (for example, Smith Keith and Smith, Keith, B.); these kinds of inconsistencies were clustered, but still, results could not be entirely precise.
The United States has 56% of patent applications and 34% of scientific documents registered. These results do not necessarily display that they produce most of the knowledge on this topic, but because of the language, most of the search engines are capable of accessing the data, unlike languages such as Spanish, Chinese, or languages spoken in India. Therefore, some designs could remain undiscovered for this investigation.
Based on the obtained results, it can be established that for this study, the effectiveness per search engine is as follows: Derwent 8.4%, Google patents 0.34%, Patentscope 3.2%, The Lens 19.9%, and Espacenet 13.95%.
The classification of the 208 prosthesis patents related to prostheses designs was obtained according to the main technology used; results show that the ESAR mechanical prosthesis is the main patent object by 44%, although claims are different for each one. All of them can be classified based on the five ESAR categories presented in this document. Outcomes also show a tendency for the use of ESAR regardless of the technology used. For 151 removable foot/ankle patent prostheses analyzed, 53% use only ESAR-type prosthesis, and 90% use ESAR in its components. From these, the more commonly used were selected and compared using Ansys, with no major differences between A and C, but for B, results show a more elastic foot thanks to its mass-centered design.
The significant trend in the use of ESAR prostheses may be because of their lower cost and greater energy efficiency. Different designs are used according to the user's lifestyle.
The minimum amount of components found for designing an active prosthesis is a DC motor, housing, a power transmission unit, a composite foot or equivalent, an energy storage device (springs, locking systems), a linkage system, an energy power supply, and a prosthesis/socket connector. From these components, most prostheses use a Maxon ® Brushless motor between 12 and 200 W. Power variations are mostly due to the gear ratio used (the more power, the lower the gear ratio), springs with stiffness between 60-445 kNm, and a Li-ion battery between 12-24 V. From these components it is especially important to consider when designing a BKA prosthesis the linkage system that needs to support most of the necessary loads, and it must be capable of tolerating at least 2 kN (for an 80 kg patient) without any failure.
Materials also play a vital role in supporting loads with 4000/5000 duty cycles per day; that is why aluminum, carbon fiber, and other composites are used in fabrication, and sometimes load reduction along the system is necessary and archived using a parallel spring arrangement.
The current development of batteries allows active prostheses to obtain enough power and charge duration without adding extra mass and weight, but for hydraulic and pneumatic prostheses, power supply currently is a problem because most of these systems are connected externally and the mass could reach up to 15 kg. Nevertheless, these systems are more efficient in mimicking human ankle movements.
For BKA prostheses, continuous growth in the development of active ones is estimated. Even though actual prostheses are capable of emulating three degrees of freedom, there is space for a complete body-integrated ankle/foot prosthesis. Funding: The present article used free software except for Derwent analytics for the patent analysis; this software was provided by the Autonomous University of Mexico State. This research was funded by CONACYT (Consejo Nacional de Ciencia y Tecnología), Grant 1009402.
Institutional Review Board Statement: Not applicable, because of this review did not involving humans or animals.
Informed Consent Statement: Not applicable, because of this review did not involving humans or animals.

Conflicts of Interest:
The authors declare no conflict of interest.