Sensorized glove and corresponding method for ergonomic analysis of the hand, particularly the hand of a worker

Abstract

A sensorized glove (2), specifically designed for ergonomic analysis, is described herein, including: an inner glove (10) comprising a plurality of extensometer sensors (ext) configured to detect relative movements between parts of a worker's hand; and an outer glove (12) comprising a plurality of pressure sensors (ps) distributed over a palmar surface and configured to detect the pressure exerted on corresponding areas of the palmar surface. Furthermore, a corresponding method for determining the type of gripping act performed by a worker is described.

Description

1 / 40 “SENSORIZED GLOVE AND CORRESPONDING METHOD FOR ERGONOMIC ANALYSIS OF THE HAND, IN PARTICULAR THE HAND OF A WORKER” FIELD OF THE INVENTION

[0001] The present invention relates to systems for ergonomic analysis of a subject, in particular a worker, and more particularly to systems applied to the analysis of movements, actions and posture of a line operator responsible for assembling vehicles. PREVIOUS TECHNIQUE AND GENERAL TECHNICAL PROBLEM

[0002] In the metal manufacturing and engineering industry, workers are called upon to perform sequential and repetitive operations, for which it is necessary to ensure a minimum level of ergonomic compatibility with humans in order to avoid risks to the health and physical integrity of operators.

[0003] In this context, the metal engineering industry, and in particular the automotive industry, is characterized by a plurality of tasks that are very different from each other and are performed numerous times during the day, with the times determined by production needs. It therefore becomes crucial to carry out a careful ergonomic assessment of these actions in order to make – if and where necessary – postural corrections and / or movement and / or action corrections.

[0004] The ergonomic analysis that can be performed using the systems and methods based on the previous technique generally explores the following techniques: - Visual analysis of the worker performed by an operator responsible for ergonomic monitoring; this process is based on capturing images of the worker at work and subsequent analysis by the operator; the process is largely performed manually / empirically and is therefore affected by subjective variability, which, although controllable, cannot be completely eliminated; it is also very demanding in terms of time resources; and - capturing the positions of body parts and / or long bones through images and video clips and processing the data thus obtained; although this methodology is practically free from subjective variability, it is significantly penalized by the fact that only part of the human body is captured. Petition 870250078613, dated 03 / 09 / 2025, page 12 / 101 2 / 40 covered and / or by the fact that the reading resolution (for ergonomic purposes) is too low; this means that the methodology is a long way from meeting the needs related to the estimation of the main ergonomic indicators.

[0005] Finally, and partly as a consequence of the above, with the systems and methods currently known any ergonomic analysis activity performed in the worker's hand is substantially impractical and it is even less possible to implement this analysis in real time and with an objective means not linked to the subjectivity of the operator responsible for ergonomic monitoring. OBJECTIVE OF THE INVENTION

[0006] The objective of the present invention is to solve the aforementioned technical problem.

[0007] In particular, an objective of the invention comprises providing a methodological support tool for ergonomic analysis at the observation and design stage that will be both objective and rapidly implementable. Specific reference will be made to the development of a methodology for analyzing the activities performed by a worker at a workstation, deriving from this analysis factors that are useful for improving product / process parameters in assembly lines at the design stage and / or optimizing workstation characteristics.

[0008] Furthermore, an objective of the present invention may be to provide a single interface for collecting ergonomic data from a plurality of devices forming part of the system, for example, displays, devices for capturing body movement, and / or devices for capturing hand movement and force (e.g., accelerometers, pressure sensors, etc.).

[0009] Additional objectives of the invention include: - Identify critical actions during a work activity; and / or - identify the main aspects that are critical from an ergonomic point of view; and / or - provide an ergonomic assessment in accordance with company and international standards; and / or Petition 870250078613, dated 03 / 09 / 2025, page 13 / 101 3 / 40 - To provide a system and method for collecting postural and dynamic data with high repeatability, and for subsequent analysis of the same applied to a worker's hand. SUMMARY OF THE INVENTION

[0010] The object of the invention is achieved by a sensorized glove and by a method having the characteristics that form the subject of the following claims, which constitute an integral part of the technical teaching provided herein in relation to the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The invention will now be described with reference to the accompanying drawings, which are provided purely as a non-limiting example.

[0012] Figure 1 is a block diagram provided by way of example of a system according to an embodiment of the invention.

[0013] Figure 2 is an overview of a device for collecting force and position data that can be applied to a worker's hand, in particular a line operator.

[0014] Figure 2A illustrates a calibration position of the device in Figure 2.

[0015] Figure 3 is a top plan view of a component of the device in Figure 2.

[0016] Figure 3A illustrates a correlation between an arrangement of sensor elements in the device of Figure 3A and a reference human anatomy.

[0017] Figure 4 is a view according to arrow IV in Figure 2.

[0018] Figure 5 is a view along arrow V in Figure 2.

[0019] Figure 6 is a cross-sectional view provided by way of example of the device in Figure 2, in particular taken along an axis of a finger.

[0020] Figures 7 and 8 are functional diagrams that illustrate the premises of a method according to the invention.

[0021] Figures 9 to 12 illustrate results provided by way of example that can be obtained during operation of the device in Figure 2.

[0022] Figure 13 is a diagram that illustrates four conditions that can be Petition 870250078613, dated 03 / 09 / 2025, page 14 / 101 4 / 40 detected using the device in Figure 2 and the method according to the invention.

[0023] Figure 14 is a schematic representation of an additional element of the system according to the invention.

[0024] Figure 15 illustrates a preferred embodiment of the element represented schematically in Figure 14.

[0025] Figures 16 to 19 illustrate some interface screens of a software for managing the system according to the invention.

[0026] Figures 20 and 21, each including a portion “A” and a portion “B”, illustrate a method of capturing a sequence of images of a task performed by the worker. DETAILED DESCRIPTION OF EMBODIMENTS OF THE INVENTION

[0027] With reference to figure 1, the number 1 designates as a whole a system according to the invention for analyzing the movements of the human body or parts thereof, in particular for analyzing the movements of a worker during the execution of the tasks assigned to him, and more particularly for analyzing the movements of an assembly line worker in the automotive sector.

[0028] System 1 comprises: - at least one device 2 (figure 1A), preferably provided as a wearable sensorized glove, to detect movements made by a worker's hand and forces applied in this way; - a wearable network of 4 sensors (figure 1B), preferably inertial sensors, where the sensors are located in the network so as to be associated with corresponding joints of the human body; preferably, the wearable network is incorporated into a garment or kit of garments such as a pair of jumpsuits; - an image capture system 6 (figure 1C) including one or more cameras configured to capture images of the worker in the work area for subsequent ergonomic assessment purposes; if multiple cameras are provided, different cameras may be associated with different filming angles; and Petition 870250078613, dated 03 / 09 / 2025, page 15 / 101 5 / 40 - a processing unit 8 configured to receive data and / or signals from the sensorized glove 2 and / or from the wearable sensor network 4 and / or from the image capture system 6 and configured to process the data and / or signals to estimate ergonomic indicators and / or obtain local information on efforts made and / or posture.

[0029] With reference to Figure 2, one embodiment of the device 2 according to the invention is illustrated therein. The device is provided as a double-walled sensorized glove. In particular, the glove 2 includes an inner glove 10 (Figure 3) and an outer glove 12 (Figures 2, 2A, 4, 5 - in all cases the presence of the glove 10 is implied).

[0030] With reference to figure 3, the inner glove 10 (which may be, for example, a commercially available sensorized glove under the trade name Cyberglove®) is configured to detect and transmit to the processing unit 6 data on the position and relative movements of the various parts of the human hand: carpus, metacarpus and phalanges. For this purpose, the glove 10 includes a first plurality of pockets 14, which are sewn inside the glove 10 so as to more faithfully follow the movements of the hand without running the risk of following the deformations of the glove, preferably have a rectangular shape and housing inside the same corresponding linear extensometers EXT (figure 6).

[0031] A second plurality of pockets 16, which are substantially U-shaped, are instead sewn onto the outer surface of the glove 10 and house corresponding linear extensometers, which, due to the U-shape of the pockets, are able to detect the relative movements of the various parts of the hand, for example, in the palmar plane (e.g., finger divarication movements).

[0032] Figure 3A illustrates an example of vector mapping that can be used to interface with data provided by strain gauges installed on glove 10. This vector map, as will be seen below with reference to Figure 8, is the basis of a method for detecting the movements of the worker's hand according to the invention.

[0033] The vector map in Figure 3A also shows a local system of Petition 870250078613, dated 03 / 09 / 2025, page 16 / 101 6 / 40 axes which can preferably be adopted to reference the motion coordinates detected by the extensometers of the glove 10. In particular, the XZ plane is a palmar plane, the XZ plane is a transverse plane, while the YZ plane is a sagittal plane.

[0034] With reference to figures 2, 4 and 5, the outer glove 12 is made of textile material, for example, elastic cotton (preferably, like glove 10) both on the palm - figure 4 - and on the back - figure 5. This makes the coupling of the outer glove 12 with the inner glove 10 more stable and precise. A second layer 18 is applied to the palm 4, made of anti-slip material designed to improve grip, for example Alcantara®.

[0035] Sewn (or otherwise applied) to the (outer) layer 18 is a plurality of pockets 20 of a preferably square shape configured to house a corresponding PS pressure sensor (see figure 4A). To facilitate the insertion of the sensor – which is substantially provided on a film holder – each pocket 20 is not sewn on at least one side (preferably it is sewn only on three sides when the pocket is square). In addition, a polyamide layer is preferably provided on the outer glove portion 12 corresponding to the palm, under layer 18. The polyamide layer facilitates the insertion of the sensors into the pockets 20.

[0036] Each pressure sensor is part of a network of 22 sensors having a layout such as to allow application of the individual PS sensors in predetermined areas of the hand. An example of such a sensor network is illustrated in Figure 4A (in a substantially open, tree-shaped configuration), and may preferably correspond to a commercially available product under the trade name Tekscan®; it is, moreover, partially visible in Figure 2. The sensor network 22 includes a plurality of electrical connections, to which the PS pressure sensors are connected and where all electrical connections lead to an interface unit 24, which allows data exchange with the processing unit 8 or in any case with any purpose-configured processing unit. Petition 870250078613, dated 03 / 09 / 2025, page 17 / 101 7 / 40

[0037] To ensure a more orderly and rational routing of the electrical connections that are part of the sensor network 22, applied to the outer glove 12, particularly on the back, in the finger area, there are one or more straps 26. The straps 26 are preferably made of self-adhesive material, such as Velcro®, and retain the electrical connections of the network 22, preferably routing them along the fingers of the outer glove 12. According to an advantageous aspect of the present invention, the straps 26 can extend circumferentially to an opening of one of the pockets 20 in such a way as to obstruct it when they are positioned on the outer glove 12. This reduces or eliminates the possibility of accidental removal of the pressure sensors from the network 22.

[0038] Again preferably, the outer glove 12 may be fitted with inserts made of an elastic band in positions corresponding to the fingers, oriented in a longitudinal direction of the fingers themselves. This allows for better adaptation to a range of different finger thicknesses.

[0039] The back (dorsal surface) of the outer glove 12 is additionally provided with a fastening portion made of self-adhesive material - for example, Velcro - whereby it is possible to quickly attach an inertial sensor 28, which makes it possible to provide the glove 2 with an absolute position reference in the case where the glove is used in the structure of a more complex system, such as one that will be described with reference to the following figures 14 and 15.

[0040] Interface 24 is secured with a strap on the worker's wrist, as illustrated in figure 2.

[0041] Referring again to figure 3, superimposed on the representation of the inner glove 10 is the map of the pockets 20 (illustrated with a double dashed and dotted line) in order to highlight the relative position between the linear extensometers EXT in pockets 14, 16 and the pressure sensors PS in pockets 20.

[0042] What follows is a description of a preferred integration method between the position / motion data provided by the sensors (EXT strain gauges) in the inner sleeve 10 and the pressure data provided by the PS sensors in the sleeve 12. It is possible to establish communication between the sensors in sleeve 10 and sleeve 12 using a protocol. Petition 870250078613, dated 03 / 09 / 2025, page 18 / 101 8 / 40 data transmission, for example, a UDP-type data transmission protocol.

[0043] The sensors in glove 10 provide as output data the angles and coordinates for twenty-two (22) hand joints identified by reference J in figure 3A. However, they are not configured to provide information on the coordinates of the fingertips. The latter is data that needs to be calculated. Calculation of fingertip coordinates

[0044] For the purposes of this calculation, the assumptions listed below are adopted (see figure 7).

[0045] i) Reference nomenclature for finger segments, starting from the metacarpals: proximal phalanx (root segment), intermediate phalanx (intermediate segment) and distal phalanx (upper segment).

[0046] It is assumed that there are five (5) distinct balls of known radius Rn. The value of radius Rn is equal to a predefined length for each distal phalanx (the lengths are estimated based on the statistically most representative dimensions of distal phalanges), n being an index ranging from 1 to 5. Each of the five balls is centered on the end of a corresponding intermediate phalanx, consequently on points TD, ID, MD, RD and PD (the first letters meaning “Thumb”, “Index finger”, “Middle finger”, “Ring finger” and “Little finger”, respectively), see figure 8.

[0047] Abduction of the distal phalanx with respect to the intermediate phalanx and abduction of the intermediate phalanx with respect to the proximal phalanx are assumed to be zero.

[0048] Referring again to Figure 7, assuming that each finger is perfectly straight, the tip of each finger (PA) is a point that belongs to the straight line passing through the two ends of the intermediate phalanx (P1 and P0) where P0 is the endpoint where the intermediate and distal phalanges meet. This condition can be written in the form of a system of equations (also satisfied by Pa) and makes it possible to express the y-z coordinates as a function of x (the system of Petition 870250078613, dated 03 / 09 / 2025, p. 19 / 101 9 / 40 reference is the same as that used for inner glove 10). (x - xo) / (xi - xo) = (z - zo) / (zi - zo) (x - xo) / (xi - xo) = (y - yo) / (yi - yo)

[0049] Pa also belongs to the ball centered at Po and having radius R, which is a known (imposed) quantity and corresponds to the length of the distal phalanx.

[0050] Physiologically, the distal phalanx Pa-Po is shorter than the intermediate phalanx P1-P0; consequently, we can assume R < L R = dist(PA,Po) = [(xa - xo)2 + (yA - yo)2 + (za - zo)2]i / 2 L = dist(Pi,Po) = [(xi - xo)2+ (yi - yo)2+ (zi - zo)2]i / 2 Note: Of the two solutions, the one with the positive sign should be chosen, since we are dealing with lengths.

[0051] By crossing the aforementioned straight line (passing through the two ends of the intermediate phalanx) with the ball of radius R centered at Po, two points are found, one of which is point Pa.

[0052] Expressing, using the straight-line equations given above, yez as a function of x (x - xo)2[1 + (yi - yo)2 / (xi - xo)2+ (zi - zo)2 / (xi - xo)2] = R2

[0053] We obtain the solutions xai and xa2: xai = xo + R / L and xa2 = xo - R / L

[0054] The correct solution will be one that satisfies the following condition: dist(PA,Pi) > L where Pai = P(xai, yAi, zai) = (xo + R / L, yo + (yi - yo)(xAi - xo) / (xi - xo), zo + (zi - zo)(xai - xo) / (xi - xo)) Pa2 = P(xa2, yA2, za2) = (xo - R / L, yo + (yi - yo)(xA2 - xo) / (xi - xo), zo + (zi - zo)(xa2 - xo) / (xi - xo))

[0055] Since Pa is known, from trigonometry we obtain dist(PA,PB) = 2Rsin(e / 2) = Ri where β is the bending angle provided by the corresponding extensometer EXT. Petition 870250078613, dated 03 / 09 / 2025, p. 20 / 101 10 / 40

[0056] To find the coordinates of Pb, the ball centered at Pa with radius equal to the distance between Pa and Pb, dist(PA, Pb) is considered: [(XB - xa)2+ (yB - yA)2 + (ZB - ZA)2] = dist2(Pa,Pb) = Ri2

[0057] The formula above can also be rewritten to obtain the coordinates of Pb = (xb· yB, zb) as follows: [(XB - X0) - (XA - X0)]2+ [(yB - y0) - (yA - y0)]2+ [(ZB - Z0) - (ZA - Z0)]2= R12 [(xb - xo) - (x - xo)]2+ [(yB - i) - (yA - i)]2+ [(zb - zo) - (za - zo)]2= Ri2(xb - xo)2- 2(xb - xo)(xa - xo) - i) + (yA - i)2+ (zb - zo)2- 2(zb - zo)(za - zo) + (za - zo)2= Ri2dist2(Pb,Po) + dist2(Pa,Po) - 2[(xb - xo) (xa - xo) + (yz - i - zo) +zo) Ri2 R2+ R2- 2xb(xa - xo) + 2xo(xa - xo) -2yB(yA - i) + 2yo(yA - i) 2zb(za - Zo) + 2zo(za - Zo) = Ri2= 4R2sin2(e / 2)

[0058] We impose that Pb is located in the plane π (ax + by + cz + d = o) passing through two points Po, Pi, P2, we fear a = -2xb(x - Xo) => XB = to / 2(x - x) b = -2yB(yA - i) => yB = b / 2(i - yA) c = -2zb(za - zo) => ZB = c / 2(zo - z) of waves Pb = (xb, yB, zb) = (a / 2(x − x), b / 2(I − yA), c / 2(zo − z))

[0059] This is an approximate solution, above all with regard to the definition of Pa and Pb for the thumb, but for the purposes of the calculation in question it is considered acceptable. Calculation of force vectors in the hand.

[0060] It is necessary first to define the reference direction and sense for the force vectors, which are individually assumed to be applied at the midpoint of each phalanx or - in the case of the palm - at the points associated with vectors represented by arrows delimited in figure 8.

[0061] To define the direction, the intersection between the plane Π passing through the three points of the three phalanges (for example, the point IM between the metacarpal and the phalanx) Petition 870250078613, dated 03 / 09 / 2025, page 21 / 101 11 / 40 proximal, the IP point between the proximal phalanx and the intermediate phalanx, and the ID point between the intermediate phalanx and the distal phalanx) and the plane Π' passing through each phalanx at the midpoint and orthogonal to Π is sought (this suggests the condition of orthogonality to the phalanx: given that the phalanx is not flexible and is contiguous to the other phalanges, the plane orthogonal to the plane containing the phalanx is orthogonal to the phalanx).

[0062] Assuming that Π passes through the points IM, IP and ID, the equation of the plane (ax + by +cz + d = 0) becomes (x - XlM)·[(ylP - yIM)-(ZIP- ZIM) - (yiD - yiM)'(ZlP- zim)] + (y - y-M) zim) - (xid - χιμ)·(ζιρ- zim)] + + (z - ζιμ)·[(χιρ - xiM)-(yiD - yiM) - (xid - xiM)-(yip - yiM)] = 0 of wave a = (yip - yiM) / zip-D zim) - zim - b) (xip - xiM^ZlD- zim) - (xid - ΧΙμ)·(Ζ|Ρ- Zim) c = (xip - xim)·(yiD - yiM) - (xid - xim)·(yip - yiM) d = - xw[(yip - yiMy(zip- zim) - (yiD - yiM- zip- zim) xiM^ZlD- zim) - (xiD - ximXzip- zim)] + zim·[(xip - xim)·(yiD - yiM) - (xid - xim)·(yip - yiM)]

[0063] To determine the Second plane Π' orthogonal to the Π plane, it is sufficient to impose the orthogonality condition for the two vectors identifying each plane (normal vectors, that is, their scalar product must be zero) and the passage of the second plane Π' + to b' + b' + 0 through the midpoint of the phalanx in question (for example, the iMM point).

[0064] A vector that satisfies the orthogonality condition for the vector (a, b, c) can also belong to the previously defined plane, and consequently the phalanx segment (for example, IP-IM) can also constitute a vector: IP-IM = [(xip - xim), (yip - yiM), (zip - zim)]

[0065] By imposing the passage through the midpoint of the phalanx (IMM = [(Xip + Xim) / 2, (Yip +Yim) / 2, (Zip + Zim) / 2]) it is possible to write the equation of the second plane as a'(xiMM) + b'(yiMM) + c'(zimm) + d'= 0 Petition 870250078613, dated 03 / 09 / 2025, p. 22 / 101 12 / 40 (xip - χιμ)·(χιρ + xim) / 2 + (yip - yiM)·(yip + yiM) / 2 + + (zip - ζιμ)·(ζιρ + zim) / 2 + d' = 0 => a' = (xip - xim) b' = (yip - yiM) c' = (zip - zim) d' = (xim2- xip2) / 2 + (yiM2- yip2 / 2 + (zim2- zip2) / 2

[0066] By combining the equations of the two planes into a system, we obtain as a solution the straight line of intersection that identifies the direction of the vector and that passes through the midpoint of the phalanx (IMM in the example).

[0067] Since it is also necessary to identify a direction to define the force vector, it is possible to consider the point of intersection between the straight line and the ball to have a radius equal to the magnitude of the force exerted at that point (there are two such points of intersection).

[0068] This point can be defined as the South pole of the ball centered at the midpoint Pimm and having a radius R equal to the sum of the pressures detected by the PS sensors at points corresponding to the pockets 20, that is, at the points associated with the fingertips.

[0069] In Cartesian coordinates, the ball of radius R centered at Pimm will have the following equation: (x - ximm)2+ (y - yiMM)2+ (z - zimm)2= R2

[0070] The points of intersection with the straight line of intersection of the two previous planes define the ends of the two vectors between which the vector sought lies (definition of the direction of the force).

[0071] Obtaining the equations of the two planes in a system, from the first equation we obtain x = (- by - cz - d) / a, and substituting into the equation of the second plane y = (-z'x - c'z - d') / b', it is possible to express both y and x as a function of z to obtain y = z (a'c - c'a) / (b'a - ab') + (a'd - d'a) / (a'c - c'a) = Az + B x = z [(b(a'c - c'a) + c(b'a - a'b)) / a(a'b - b'a)] + [(b(d'a - a'd) + d(c'a - a'c)) / a(a'c - c'a)] = Cz Petition 870250078613, dated 03 / 09 / 2025, p. 23 / 101 13 / 40 + D where A = (a'c - c'a) / (b'a - ab') B = (a'd - d'a) / (a'c - c'a) C = [(b(a'c - c'a) + c(b'a - a'b)) / a(a'b - b'a)] D = [(b(d'a - a'd) + d(c'a - a'c)) / a(a'c - c'a)]

[0072] The above formulation should only be used in the case where aa 0 and a'c a c'a and a'b a b'a.

[0073] Substituting xey into the equation of the ball, we obtain the two coordinates zi and Z2 (belonging to the set of real numbers; in any case, it is advisable to verify that the discriminant of the equation is positive, that is, (β2- 4αγ) > 0).

[0074] Assuming α = (C2+ A2+ 1) β = 2[C(D - ximm) + A(B - yiMM) - zimm)] Y = - R2+ (D - ximm)2+ (B - yiMM)2+ zimm2we obtain zi = [-β - (β2- 4αγ)1 / 2] / 2α Z2 = [β - (β2- 4αγ)1 / 2] / 2α where Pi = P(xi,yi,zi) = (Cf-β - (β2- 4αγ)1 / 2] / 2α + D, A[-β (β2- 4αγ)1 / 2] / 2α + B, [-β - (β2- 4αγ)1 / 2] / 2α) P2 = P(x2,y2,z2) = (C[+β - (β2- 4αγ)1 / 2] / 2α + D, A [+β (β2- 4αγ)1 / 2] / 2α + B, [+β - (β2- 4αγ)1 / 2] / 2α)

[0075] The correct triad will be the one with the shortest distance from the coordinate corresponding to the tip of the thumb Ptx, namely PiMMF(x,y,z) = min (dist(Pi, Ptx), dist(P2, Ptx)) where: dist(Pi, Ptx) = [(xi - ximmf)2+ (yi - yiMMF)2+ (zi - zimmf)2]1 / 2e dist(P2, Ptx) = [(x2 - ximmf)2+ (y2 - yiMMF)2+ (z2 - zimmf)2]i / 2 Petition 870250078613, dated 03 / 09 / 2025, p. 24 / 101 14 / 40

[0076] For it to be a Cartesian plane, all the coefficients that identify (vector orthogonal to the plane) can never be simultaneously zero, and consequently we can never have a = b = c = 0 simultaneously.

[0077] If a = 0, only the following interesting and non-degenerate cases can be found: a = b = 0 eca 0 a = 0 eba 0 ec of any value

[0078] With a = b = 0 and a ≠ 0, we have z = -d / c, and the two subcases with b' ≠ a ≠ 0 and b' = 0 will be possible.

[0079] If a = b = 0 and a ≠ 0, substituting z into Π' makes it possible to obtain y as a function of x (or x as a function of y) depending on whether a' or b' are possibly zero.

[0080] If a' = 0 and b' ≠ 0, we obtain y = (c'd - d'c) / cb' and z = -d / c, which, after being substituted into the equation of the ball, will provide the solutions xi and X2: xi = xiMM - [(R2- (-d / c - zimm)2- ((c'd - cd') / cb'- yiMM)2]1 / 2 X2 = ximm + [(R2- (-d / c - zimm)2- ((c'd - cd') / cb'- yiMM)2]1 / 2where Pi = P(xi,yi,zi) = (ximm - [(R2- (-d / c - zimm)2- ((c'd - cd') / cb'- yiMM)2]1 / 2, (c'd - cd') / cb', d / c) P2 = P(x2,y2,z2) = (ximm + [(R2- (-d / c - zimm)2- ((c'd - cd') / cb'- yiMM)2]1 / 2, (c'd - cd') / cb', d / c)

[0081] The correct triad will be one with the shortest distance from the coordinate corresponding to the tip of the thumb Ptx, namely PiMMF(x,y,z) = min (dist(Pi, Ptx), dist(P2, Ptx)) where: dist(Pi, Ptx) = [(xi - ximmf)2+ (yi - yiMMF)2+ (zi - zimmf)2]1 / 2e dist(P2, Ptx) = [(x2 - ximmf)2+ (y2 - yiMMF)2+ (z2 - zimmf)2]1 / 2

[0082] If a' Ψ 0 and b' = 0, we obtain x = (c'd - d'c) / a'c and z = -d / c Petition 870250078613, dated 03 / 09 / 2025, p. 25 / 101 15 / 40 which, when substituted into the ball equation, yields the solutions yi and y2: yi = yiMM - [(R2- (-d / c - zimm)2- ((d'c - cd') / ca'- ximm)2]1 / 2 y2 = yIMM + [(R2- (-d / c - zimm)2- ((d'c - cd') / ca'- ximm)2]1 / 2 and consequently Pi = P(x1,y1,z1) = (yiMM - [(R2- (-d / c - zimm)2- ((d'c - cd') / ca'- ximm)2]1 / 2, (c'd - d'c) / a'c, - d / c) P2 = P(x2,y2,z2) = (yiMM + [(R2- (- d / c - zimm)2- ((d'c - cd') / ca'- ximm)2]1 / 2, (c'd - d'c) / a'c, d / c)

[0083] The correct triad will be the one that has the shortest distance from the coordinate corresponding to the tip of the thumb Ptx, namely Pimmf (x,y,z) = min (dist(P1, Ptx), dist(P2, Ptx)) where: dist(P1, Ptx) = [(xi - ximmf)2+ (y1 - yiMMF)2+ (zi - zimmf)2]1 / 2e dist(P2, Ptx) = [(x2 - ximmf)2+ (y2 - yiMMF)2+ (z2 - zimmf)2]1 / 2

[0084] After this calculation, the average points of application of the forces in the various areas of the hand are consequently known, which are derived from the readings taken by the PS sensors, in relation to the positions of the J joints, the postural data of which are collected by the EXT extensometers.

[0085] This allows the direction of the force itself to be defined.

[0086] The operation of glove 2 is described below.

[0087] Glove 2 can be used to obtain postural information on the hand (when inertial sensor 28 is present in the glove), to obtain the direction of the resultant force when associated with the reading of pressure intensities with the corresponding orientation of the surface on which the pressure acts (direction of the force on the various sensors) and mainly to recognize the type of gripping action performed by the worker when the finger postures are associated with the pressure exerted by the fingers most involved in the posture being detected.

[0088] In particular, four different types of holding acts can be detected: Petition 870250078613, dated 03 / 09 / 2025, page 26 / 101 16 / 40 a) GRIP: this is the type of holding action performed, for example, when gripping an angle torque wrench set in a position corresponding to a joint; for the purposes of the test, the results of which appear in Figure 9, the holding action was performed simulating the screwing operation for a duration of 3 s starting from a neutral posture (where "neutral posture" means a hand at rest, that is, in a posture without any type of holding action and without any load detected by the sensors); b) PALMAR: this is the type of gripping action performed, for example, when lifting a standard weight; for the purposes of the test, the results of which appear in Figure 10, a standard weight of 2 kg was chosen, which (starting from a neutral posture) is picked up from the workbench and lifted, and this posture is held for 3 s, after which the weight is repositioned on the workbench; c) GRASPING: This is the type of gripping action performed, for example, when lifting a bucket with a molded handle of known weight; for the purposes of the test, the results of which appear in Figure 11, a bucket weighing 5.8 kg was chosen; the bucket is picked up from the ground (starting from a neutral posture) until the arm is fully extended, the posture is held for 3 seconds, and the bucket is placed back on the ground; and d) TIGHTENING: this is the type of holding action performed, for example, when a screw is picked up from the work surface; for the purposes of the test whose results appear in Figure 12, a sequence was considered that comprises (starting from the neutral posture): picking up a screw from the work surface; positioning the screw in the nut; and manually screwing the screw into the nut with five turns of tightening.

[0089] To collect information for the purpose of recognizing the types of holding acts, calibration tests are conducted on a plurality of operators.

[0090] In a case provided by way of example, calibration was performed on twelve operators according to the following operating scheme:

[0091] Assessment of each operator's maximum strength to evaluate force ranges; this involved the use of instruments such as a gripper and a grip gauge. Petition 870250078613, dated 03 / 09 / 2025, page 27 / 101 17 / 40 until maximum strength, with three repetitions;

[0092] Evaluation of the accuracy and precision in capturing force data (deviation from the expected value and the average value) by gripping the handle with a force of 5 kg maintained constant for 3 s, with five repetitions; and

[0093] Assessment for recognition of the type of holding act (grasping / squeezing / palming / hooking); a cycle of the four types of holding act was performed in succession (with a neutral posture assumed between each type of holding act), with three repetitions.

[0094] The assessment of the maximum force exerted by each operator in both the case of TIGHTENING (using the tightening gauge) and in the case of GRIPING (using the gripper) - cf. point 1 above - allows: - assessment of the range of forces in which calibration tests and their results (and the results of future recordings in general) are statistically located; - assessment of the variability in the ability to exert force among the various subjects in the available sample for each type of gripping act; and - Defining a minimum limit value, lower than the maximum strength of the weakest operator, to be associated with the reference used in the software logic to stop recognizing the types of holding actions.

[0095] If in the sample the “weakest” operator is able to exceed the minimum threshold, then there is reasonable certainty that it is possible to identify a given type of holding act for all components of the sample; that is, if the minimum threshold is reached by the weakest subject, certainly the strongest will be able to reach and exceed (even abundantly) the activation value. The minimum threshold activates the recognition of the type of holding act, filtering out any possible background noise or sensor activations that should not be attributed to the type of holding act.

[0096] Then, the pressure maps, captured by the PS sensors, were analyzed in relation to the various postures of all twelve subjects to identify the areas of the hand involved in the various types of gripping action. Petition 870250078613, dated 03 / 09 / 2025, page 28 / 101 18 / 40

[0097] The pressure / force values ​​exerted by the various subjects during the various types of holding acts were evaluated to assess the variability of the values ​​due to the subjectivity of test execution.

[0098] The results obtained from the pressure maps captured by the PS sensors highlighted for the various postures the part involved in the pressures that allow the definition of the various types of holding action, meaning that a given configuration of postures and pressures allows (instrumental) identification of a specific type of holding action and consequently definition of the reference condition of the same (for the instrument) to automatically recognize which type of holding action is being performed. Type of holding action “by GRASPING” - figures 9A and 9B

[0099] The pressure map in Figure 9A shows widespread involvement of the entire hand in applying pressure to the handpiece. With reference to the diagram in Figure 9B, some variability can be observed in the force values ​​detected for the various operators. The diagram in Figure 9B (this also applies to subsequent figures 10B, 11B, 12B) graphically represents the tracing of the average force applied to the palm of the hand and detected by the pressure sensors (whose topographic map can be seen in Figure 9A, with relative activation intensity at the base of the shading). The average pressures of each sensor multiplied by the area involved in the detection provide an average force value (F), and the average of the forces (F) detected by the various sensors is the force F represented by the time graph. This is data provided by the instrument for capturing the pressures. Type of hand grip action “PALMAR” - figures 10A and 10B

[0100] The pressure map in Figure 10A shows that the areas most involved in the pressure exercise are the middle and distal phalanges of all five fingers. It is generally observed that the noise in the rest of the hand is no higher than 20-30% of the total force exerted by the hand.

[0101] With reference to the diagram in figure 10B, a certain variability in the force values ​​detected for the various operators can be observed. Type of holding action “BY HOOKING” - figures 11A and 11B Petition 870250078613, dated 03 / 09 / 2025, page 29 / 101 19 / 40

[0102] The pressure map in Figure 11A shows that the areas most involved in the pressure exercise are the middle and distal phalanges of the four fingers and the upper part of the palm. The noise on the rest of the hand is no higher than 20% of the total force exerted by the hand.

[0103] With reference to the diagram in figure 11B, a certain variability can be observed in the force values ​​detected for the various operators. Type of holding action “BY SQUEEZING” - figures 12A and 12B

[0104] The pressure map in Figure 12A shows that the areas most involved in the pressure exercise are the middle and distal phalanges of the thumb, index finger, and middle finger. Noise in the rest of the hand is always present and is generally approximately 30-40% of the total force exerted by the hand.

[0105] With reference to the diagram in figure 12B, a certain variability can be observed in the force values ​​detected for the various operators. Logic for recognizing the type of act of holding.

[0106] For each type of gripping action, based on the analysis of previous pressure maps, a given area of ​​the hand was selected where most of the pressure exerted by the hand itself should be concentrated.

[0107] To identify the type of gripping action, the pressures and positions of the hand areas involved in the gripping action were analyzed and, through verification, the pressures and postures of the remaining hand area were also analyzed. This is illustrated in figures 9C, 10C, 11C and 12C, where in particular the areas where a pressure measurement higher than a limit value is expected to be found are circled.

[0108] Instead, the areas of the pressure map where, for each type of holding action, the noise concentration (where applicable) is expected are covered with dashed outlines. The pressure map visible in each figure of this group of figures is arranged as a set of matrices each associated with a PS sensor in a pocket 20, where each cell of each matrix (the so-called sense1, designated by SS) contains the value detected by a sensitive element of the PS sensor in the corresponding area of ​​influence. Petition 870250078613, dated 03 / 09 / 2025, page 30 / 101 20 / 40 The act of holding onto something by grabbing it.

[0109] With reference to figure 9C, for the act of gripping, pressure must be present on all seventeen areas of the hand. All SS senses are involved in the act of gripping, i.e., 361 senses in the specific example illustrated here. The total pressure exerted in this type of gripping act must be at least 30% higher than the pressure value in the neutral posture. In most cases, the pressure is concentrated in the finger area and the lower outer area of ​​the palm (L-shaped sensor). The hand posture must correspond to the gripping act. The act of holding a palm.

[0110] For the act of holding the PALMAR grip, pressure should be concentrated primarily on the middle and distal phalanges of all five fingers. There should be no contact with the palm area. The sense1s SS in question for the act of holding the PALMAR grip in this specific example are 140 in number (38% of the total area). In this area, the pressure should be at least 51% of the total pressure.

[0111] There may be residual pressure (noise) in the rest of the hand, which, however, should not exceed 49% of the total pressure. The abduction angles between the four fingers should be more than 3° to differentiate them from those of the act of gripping, and the hand posture should correspond to the act of gripping. The act of holding onto something by hooking it.

[0112] For the act of gripping by hooking, pressure should be concentrated on the proximal and middle phalanges of four fingers and on the upper part of the palm (pulley). The SS senses in question for the act of gripping by hooking in this specific example are 172 in number (48% of the total area). In this area, the pressure should be at least 51% of the total pressure. There should be no pressure on the thumb; that is, any possible noise present in the thumb should not exceed the value acquired in the neutral posture. In the rest of the hand (and in particular in the distal phalanges) there may be residual pressure, which, however, should not exceed 49% of the total pressure. The hand posture should correspond to the act of gripping. The act of holding by squeezing. Petition 870250078613, dated 03 / 09 / 2025, page 31 / 101 21 / 40

[0113] For the act of gripping with three fingers, the pressure should be concentrated mainly on the middle and distal phalanges of the thumb, index finger, and middle finger. The SS senses in question for the act of gripping with three fingers in this specific example are 84 in number (23% of the total area). In this area, the pressure should be at least 51% of the total pressure. There should always be pressure on the thumb. In the rest of the hand, there is often residual pressure (noise) due to hand flexion, which should not exceed 49% of the total pressure. The abduction angle between the index and middle fingers should be less than 5°. The hand posture should correspond to the act of gripping.

[0114] The angles describing hand posture were evaluated during the tests thanks to data collected through the EXT extensometers in the inner glove 10. A certain repeatability was detected in the hand joint angles in the various repetitions of the intra-subject test. As an example, figure 13 illustrates the angle captured by the MPJ joint, which is marked with the same reference as that used in figure 3 and corresponds to the point where the middle finger meets the palm.

[0115] In summary, using the sensorized glove 2, it is possible to provide a method for ergonomic analysis of a worker's hand that comprises the following steps: - receive data from the first sensor of the plurality of EXT extensometer sensors of the inner sleeve 10; - to associate - as described previously - the data from the first sensor to a vector map of the hand, comprising a plurality of nodes associated with corresponding joints of the hand and a plurality of segments that connect the nodes to each other, where one or more EXT strain gauge sensors associated with a node are configured to detect a relative angular position between a pair of segments connected to the node, each data point from the first sensor comprising a relative angular position between the pair of segments connected to the corresponding node; - receive data from the second sensor from the plurality of PS pressure sensors of the outer sleeve 12, each data from the second sensor comprising information about the pressure detected along an area of ​​the pressure sensor. Petition 870250078613, dated 03 / 09 / 2025, page 32 / 101 22 / 40 corresponding PS; - determine, for each PS pressure sensor, a position of a point of application of a resultant force, which is a function of the data from the corresponding second sensor, with respect to the nodes (J joints) of the vector map (see previous description); and - To determine, based on data from the first sensor in combination with data from the second sensor, the type of grasping action performed by the hand.

[0116] In particular, the method that forms part of the invention, which determines the type of grasping act performed by the hand, includes: - define a pressure map, comprising a plurality of map areas associated with corresponding pressure sensors PS of the outer sleeve 12, each map area being a set of sensitive elements SS having a respective area of ​​influence, each sensitive element SS being associated with information representing a pressure value detected in a homologous area of ​​influence on the corresponding pressure sensor PS; - Define engagement limits for each map area, where each engagement limit is defined as a fraction of SS-sensitive elements involved relative to the total number of SS-sensitive elements in the map area and is representative of a different type of holding act; - record the information provided by the sensitive SS elements of each map area and compare it with the boundaries of engagement that can be applied to the different types of holding act; in this respect, each type of holding act may include its own set of boundaries of engagement that is descriptive of the expected situation for the holding act in question; and - Determine the type of holding act based on the result of comparing the information provided by the SS sensitive element of each map area with the engagement limits that can be applied to the different types of holding acts.

[0117] Finally, it should be noted that it is possible to use the data from the first and second sensors for a cross-check on determining the type of act of Petition 870250078613, dated 03 / 09 / 2025, page 33 / 101 23 / 40 hold. In particular, the information determined based on the recording of information from the pressure map mapping areas must be compatible with the information determined based on postural data reconstructed by means of the EXT extensometer sensors of the inner sleeve 10.

[0118] With reference to figures 14 and 15, a wearable sensor network 4 that can be used in a system according to the invention is preferably provided as a sensorized pair of suit jackets (in one piece or two pieces) or an ensemble of wearable accessories 30 (e.g., belts, jackets, etc.) that individually carry one or more inertial sensors – or sensors in general – designed to provide indications about postural angles and Cartesian coordinates at predetermined points on the body. An example of such a wearable network is represented by the sensorized suit produced by Xsens Technologies BV, P.O. Box 559, 7500 AN ENSCHEDE, Netherlands.

[0119] The sensors that can be installed on the wearable network 4 include, in combination or as alternatives to each other: - accelerometers, to measure acceleration in three-dimensional space; - gyroscopes, to measure the orientation of Earth's gravity; and / or - magnetometers, to have a common reference for the system, that is, the Earth's magnetic field.

[0120] The wearable sensor network 4 provides that the sensors 30 are located on corresponding parts of the operator's body, where these parts of the operator's body have a correspondence with a human skeleton reference scheme illustrated in figure 14 and indicated by 4RIF.

[0121] The 4RIF reference scheme includes segment elements to define bones of the human skeleton connected by articulation elements to define joints of the human skeleton. The articulation elements are preferably point-like elements and are identified by the references jC1 Head, jT1 C7, jT9T8, jL1T12, jL4L3, jL5S1, Root, jRightC7 Shoulder, jRight Shoulder, jRight Elbow, jRight Wrist, jLeftC7 Shoulder, jLeft Shoulder, jLeft Elbow, jLeft Wrist, jRight Hip, jLeft Hip, jRight Knee, jRight Ankle, jRight Ball Foot, jLeft Knee, jLeft Ankle, Petition 870250078613, dated 03 / 09 / 2025, page 34 / 101 24 / 40 jLeftBallFoot. The same references are also reproduced in Figure 15. A table containing the legend follows below. Joint Name | Corresponding Anatomical Joint | jC1Head | Head Joint - C1 | jT1C7 | Sternum Joint - T1C7 | jT9T8 | Vertebral Joint (non-sensorized) | jL1T12 | Vertebral Joint (non-sensorized) | jL4L3 | Vertebral Joint (non-sensorized) | jL5S1 | Vertebral Joint (non-sensorized) | Root | Vertebral Joint (non-sensorized) | jRightC7Shoulder | Right Posterior Shoulder Joint (scapula) | jRightShoulder | Right Shoulder Joint | jRightElbow | Right Elbow Joint | jRightWrist | Right Wrist Joint | jRightHip | Right Hip Joint | jRightKnee | Right Knee Joint | jRightBallFoot | Right Foot Joint | jRightAnkle | Right Ankle Joint | jLeftC7Shoulder | Left Posterior Shoulder Joint (scapula) | jLeftShoulder | Left Shoulder Joint Left Elbow: Left elbow joint; Left Wrist: Left wrist joint; Left Hip: Left hip joint; Left Knee: Left knee jointLeft Ball Foot Joint (J): Left Ankle Joint

[0122] The inertial sensors 30 are arranged at representative points of a respective segment element; in particular, they can be fixed (by means of the aforementioned bands or other types of accessories generally usable) to respective parts of the body in positions corresponding to representative points of the segment element identified (at a software level) by the articulation elements of figure 14 that must be monitored through the network 4.

[0123] Network 4 allows the capture of a person's body kinematics (in this case, a worker). In other words, it allows the capture of trajectories, postures, and angles of rotation that each segment and joint element of the body assumes during any activity, and specifically a Petition 870250078613, dated 03 / 09 / 2025, page 35 / 101 25 / 40 work activity. Network 4 is configured to provide output values ​​representing a rotation in space (i.e., rotations around the x, y, and z axes) and the spatial coordinates for each monitored point.

[0124] The above values ​​can be captured by processing unit 8, which can automatically evaluate the postures assumed by the worker throughout the work cycle. For example, it is possible to process the coordinates of the joints to obtain rotation angles of the body joints, according to the indications specified in the ergonomics standards.

[0125] The analysis that can be performed via network 4 can be integrated with the analysis conducted on the worker's hands via glove 2. In general, it is possible to define network 4 in communication with one or more gloves 2.

[0126] The pattern distribution of the joints where the postural angles used for the purposes of ergonomic assessments are detected is schematically represented in Figure 14. The ergonomically significant angles at which the system according to the invention is designed to operate are as follows: - back angle; - shoulder angle; and - Elbow and knee angle.

[0127] For these angles, some definitions and some assumptions are introduced, namely: Vertical line: this is the line between the sagittal plane and the frontal plane; Hip line: this is the line between the jRightHip and jLeftHip points (joints); Back line: this is the line between the points (joints) jT1C7 and Xsens root; Shoulder line: this is the line between the jRightShoulder and jLeftShoulder points (joints) of Xsens; Arm line: this is the line between the jRightShoulder and jRightElbow points (joints) of Xsens; Forearm line: this is the line between the points (joints). Petition 870250078613, dated 03 / 09 / 2025, page 36 / 101 26 / 40 jRightElbow and jRightWrist; Thigh line: this is the line between the right hip and right knee points (joints); and Leg line: this is the line between the jRightKnee and jRightAnkle points (joints). Calculation / solution for back movements

[0128] Conditions: - The hip line is aligned with the horizontal and frontal planes; - The angle of back flexion is the angle between the vertical line and the back line projected onto the sagittal plane; - The lateral slope of the back is the angle between the vertical line and the back line projected onto the frontal plane; and - Back torsion is the angle between the shoulder line (measured after the back is aligned with the vertical line) and the hip line projected onto the horizontal plane. Back - forward flexion: calculation of reference geometries. Back vector calculation. Pshoulders= (jRightShou Ider + jLeftShoul der ) / 2 Pis =(jRightHip j^ftHip2——— V = P, , .- P . torso shoulders hips Hip vector calculation — Vhips= jLeftHip- jRightHi (2) (3) (4) Sagittal plane calculation Ω ——

[0129] Initially, the hip vector Vhips is projected onto the horizontal plane, namely, V^XY = V^ with the component Vhips(z)=0.

[0130] So, the sagittal plane Ω is calculated as I x Ω = (a, b, c) · < y + d I z Petition 870250078613, dated 03 / 09 / 2025, page 37 / 101 27 / 40 with: (a, b, c) = VhipXY d = VhlpSXY(0) · PhlpS(0) + VhlpXY(1) · Phlpi1) + V: · γ , · PhlpS(2) Application of solution method Projection of the back vector onto the sagittal plane

[0131] Given the following: (1, m, n) = Vtorao(a, b, c) = direction coefficients of the sagittal plane Ω (1 Ί / C mj k = I · c - a I / I b -· c I kn ) / f nJ

[0132] The direction coefficients of the plane IX y+ dpi z

[0133] Which contains the straight line passing through Vtorsosão calculated as follows: api = 1; bpi = k;cpi =- (1 + k·m) / n

[0134] Consequently, the projection of the desired back vector corresponds to the intersection of the two planes Π and Ω: V1 = (api'bpi'cpjX (a,b,c

[0135] Intersection vector between sagittal plane and horizontal plane V2 = VhipSX (0,0,1)

[0136] Calculation of forward bending as angle between v1 and v2(--3 -1 V · v2 TF = cos u|V1| · |vjJ

[0137] The proposed method is independent of the subject's position relative to the reference triad of the recording system. In addition, a check on the arccosine Petition 870250078613, dated 03 / 09 / 2025, page 38 / 101 28 / 40 allows for distinguishing the correct sign of the value in the case of flexion (positive sign) and extension (negative sign) of the back.

[0138] Schematic solution of the calculation for advanced bending Calculation of reference geometries Back vector calculation Hip vector calculation Sagittal plane calculation Application of solution method Vi = projection of the back vector onto the sagittal plane V 2 = intersection vector between the sagittal plane and the horizontal plane Calculation of forward flexion (Tf) as the angle between Vi and V2. Calculation / solution for shoulder movements.

[0139] Conditions: The shoulder line is aligned with the horizontal and sagittal planes; - The flexion-extension of the arm is the annulus between the vertical line and the arm line projected onto the sagittal plane; - Arm abduction is the angle between the vertical line and the arm line projected onto the frontal plane; - If the hand is behind the hip line, there is an extension; - If the angle between the shoulder line and the arm line, projected onto the horizontal plane, is less than 45°, there is flexion; otherwise, there is abduction. Calculation / solution for elbow and knee movements.

[0140] Conditions: - Elbow flexion-extension is the solid angle between the upper arm line and the forearm line; - Knee flexion-extension is the solid angle between the thigh line and the leg line. Calculation / solution for other movements

[0141] Since the position of the skeletal endpoints (top of the head, end of the hand, end of the foot) is not available, the Petition 870250078613, dated 03 / 09 / 2025, page 39 / 101 29 / 40 lateral movements, flexion-extension, and pronation-supination (twisting) of the hands and head are calculated based on information coming from sensors on the wearable network 4.

[0142] Hand pronation and supination are instead calculated starting from elbow rotations. Biomechanically, the wrist does not rotate: in other words, the hand does not rotate relative to the forearm (using the wrist as a joint). Hand pronation and supination occur because, starting from the elbow, there is a rotation of the ulna relative to the radius; the two long bones are always limited at the same point in both the wrist and the elbow, but they can be crossed over each other, thus generating hand pronation and supination. Consequently, hand pronation and supination are caused by “rotation” of the long bones starting from the elbow.

[0143] The objective of system 1 according to the invention is the creation of a methodological support instrument for ergonomic analysis in the observation and design stage that will be both objective and quick to use.

[0144] The development of the assisted analysis methodology is directed towards an objective and detailed evaluation of the activity performed at a workstation, by drawing, from the verification stage, the main factors useful for improving, at the design stage, the process / product parameters in assembly lines.

[0145] The overall objective is the implementation and development of methods for ergonomic process evaluation, applying the data and information obtained also in the design stage of new workstations or in the redesign / modification stage of existing workstations.

[0146] For this purpose, system 1 provides a simple interface for collecting ergonomic input data from a work activity, namely, coming from: a) the image capture system 6 (figure 1C) - avi, mpg files, etc.; b) the wearable sensor network 4; and c) the sensorized glove 2.

[0147] System 1 is consequently capable of: Petition 870250078613, dated 03 / 09 / 2025, page 40 / 101 30 / 40 - synchronize the wearable sensor network 4 and the sensorized glove 2, in particular synchronize and use together the data provided in this way; - Integrate and convert the collected data into ergonomic results; - Identify the most critical activities within a worker's set of tasks; - Identify the main critical aspects from an ergonomic point of view; and - Provide an ergonomic assessment in accordance with international (and company) standards.

[0148] Ergonomic methods currently available for assessments include: OCRA Checklist (manual and automatic) OCRA Index (manual) MURI (manual and automatic) EM-MURI (manual) RULA (manual) Sue Rodgers (manual) NIOSH (manual) Snook & Ciriello (manual)

[0149] In addition to the above, the system allows for the prior arrangement of data for transfer to specific software for ergonomic and time analysis known as TiCon (automatic).

[0150] The features of system 1 include: - guided viewing of the music video; - possibility of collecting and storing data according to a clearly defined procedure; - function to accelerate data collection (also automatically by motion capture systems, such as the network 4 and the glove 2), and organization of the storage database; - possibility of analyzing the data (extracted from the video clip) for Petition 870250078613, dated 03 / 09 / 2025, page 41 / 101 31 / 40 Identification of ergonomic risk factors; - possibility of reprocessing the observed data based on variations in fundamental ergonomic parameters; - Organization of data characterizing localized forces and pressures due to the use of tools during the work process (dynamic analysis, using the sensorized glove 2 and, to a lesser extent, using the wearable sensor network 4) and association of the above data with postures and description of the actions performed (kinematic analysis; in this case, the wearable sensor network has at least the same importance as the sensorized glove 2).

[0151] System 1 allows for two different methods of data analysis, namely: - manual analysis (from video); and - Automatic analysis (based on data from network 4 and / or glove 2).

[0152] Manual analysis allows for the execution of an assisted ergonomic analysis starting from one or two synchronized video clips, through: - Identifying critical postures by pressing specific keys on the keyboard while the video clip is playing; and - Automatic calculation of times and required ergonomic indices.

[0153] Automatic analysis allows reading data at the output of the wearable sensor network 4 and the glove 2 (which records the user's movements) for automatic recognition of most critical postures.

[0154] The above analysis was implemented primarily for some ergonomic methods (OCRA and MURI checklist); additionally, some input data is also available for the Ticon© software.

[0155] For operation of system 1 in manual mode, it is essential to capture at least one video clip of the workstation to be analyzed. If the video clip is the only type of data captured, then only manual analysis is possible. If, instead, video clip capture is combined with data capture from the wearable network 4 and / or from the sensorized glove 2, then automatic analysis is possible.

[0156] The software has two interface modes: Petition 870250078613, dated 03 / 09 / 2025, page 42 / 101 32 / 40 “input mode”, where the software autonomously activates manual or automatic mode according to the loaded files; in the case of activation in manual mode, it is possible to select the ergonomic items that must be implemented manually; and “output mode”, where visible (on a screen) is the time distribution of the manually or calculated input characteristics starting from the data acquisition files of network 4 and glove 2. System operating logic according to the invention

[0157] System 1 according to the invention is preferably fully implemented and managed through software. The prerequisites for ergonomic analysis are the availability of at least one video clip of the work activity to be analyzed, or else (in combination or as an alternative) an IVR (Immersive Virtual Reality) system.

[0158] The second option is implemented in particular at the design stage when the physical workstation for the operator does not yet exist, and provides for the integration of the glove 2 and the wearable network 4 with the IVR system in combination with the use of real tools and objects, such as wrenches, screwdrivers and line structures in a context commonly identified as “mixed reality”. To be able to exercise this option, it is necessary to have available a CAD model of the workstation that must be analyzed, and it is similarly necessary to know the work cycle of the task to be performed.

[0159] The software by which the operation of system 1 is implemented is configured to read most commercially available types of video files. In the case of manual analysis, it is possible to proceed with post-processing analysis of the various ergonomic items that should be evaluated.

[0160] In the case of operation in IVR mode, instead of the availability of the video clip of the operator working at the workstation, a video file is provided, obtained after implementing the following operating logic: - Importing the CAD model of the workstation to be analyzed into the virtual environment; Petition 870250078613, dated 03 / 09 / 2025, page 43 / 101 33 / 40 - Defining the movement logic of the tools (wrench, screwdriver, etc.) to be used in the virtual environment; - assigning the real object to the corresponding virtual object; - preparation of the user who will be using both network 4 and glove 2; and - reproduction of the work task and recording of the data.

[0161] In this way, it is possible to generate a video file (for example, in .mpg format) to be used instead of the video clip captured online at the actual workstation, integrating it with the data recorded by network 4 and glove 2.

[0162] In this regard, reference may be made to figures 20 and 21, which illustrate, in section A, a user U equipped with the wearable net 4, a pair of gloves 2 (one for the right hand and one for the left hand), and a pair of VSR glasses.

[0163] Figure 20 shows user U, who is interacting with a structure F that represents a resting surface available at the physical workstation so that the user is supported in the posture as in the real work environment. Figure 20B illustrates the virtual transposition of user U's condition in Figure 20A, which in turn corresponds to the display available to the user through the VSR glasses.

[0164] Figure 21 shows, instead, user U handling a real tool T, for example, a screwdriver. Figure 21B illustrates the virtual transposition of user U's condition in Figure 20A, which in turn corresponds to the display available to the user through the VSR glasses.

[0165] It should be noted that the use of physical elements such as the F-frame or the T-slot in the IVR system is preferable to avoid misinterpretations of postural data and, above all, of grasping act data.

[0166] Initially, the software is set to INPUT mode. In this mode, the processing unit 8 automatically captures data from the wearable sensor network 4 and / or from the sensorized glove 2, or receives input data through manual data entry. In this mode, it is possible to select the ergonomic method to be used for the evaluation.

[0167] Figure 16 shows an interface that allows opening Petition 870250078613, dated 03 / 09 / 2025, page 44 / 101 34 / 40 simultaneous recording of a video clip captured by one or more image capture devices 6 and / or a video clip generated by means of the IVR system, a reconstruction of the worker's body posture obtained through data coming from the network 4 and / or a reconstruction of the hand posture that can be obtained by processing the data received from the sensorized glove 2.

[0168] After the animation and video files of the software simulation have been opened (the software simulation being represented in the central image of figure 16 and corresponding to a human model generated in a software environment and managed in a software environment; operationally, it is the model of the human skeleton associated with the sensor network), access is obtained to the interface that allows representation of an analysis of some movements simultaneously with display, for example, of actions, pressures / forces, or types of hand gripping acts, which can be considered important for ergonomic analysis purposes. OCRA Checklist Method

[0169] The OCRA (Occupational Repetitive Actions) checklist method consists, as is known, of a synthetic index for assessing work-related risk factors that affect musculoskeletal disorders of the upper limbs, and is capable of identifying the risk of overload of the upper limbs due to repetitive work.

[0170] If network 4 data is available (i.e., if data files captured by the device are available), the following pre-calculated features are automatically available: Arm almost at shoulder height; Hand above the head; Extreme pronation and supination of the elbow; Extreme flexion-extension and wrist deviation.

[0171] If data captured by the sensorized glove 2 is also available, the following additional features are also automatically available (according to the methodologies described above): Recognition of the various types of holding acts, with notification of Petition 870250078613, dated 03 / 09 / 2025, page 45 / 101 35 / 40 presence of an incongruent gripping action (squeezing grip, palmar grip, or hooking grip: in this regard, these types of gripping actions can cause biomechanical overload if repeated frequently or even damage due to compression of the synovial capsules of the hand tendons; consequently, they should be detected and, if performed frequently, contribute to increasing the ergonomic risk index); and Recognition of static actions.

[0172] Ergonomic items that are not analyzed automatically can be analyzed manually by the user. An example of a graphical interface that can be used for this purpose is illustrated in Figure 17, section A.

[0173] Ergonomic items can be selected from an “ITEM” menu that appears on the right side of the screen. In this mode, it is possible to analyze up to four ergonomic items one at a time, but it is also possible to perform the analysis by selecting only one item at a time.

[0174] In the case where a data file generated by capture via network 4 was selected, after the user enters an enable command (for example, by pressing a “play” or “start” button, see reference PL in figure 17; reference STP identifies a stop key instead), the software automatically calculates the ergonomic information of the OCRA Checklist method for the items mentioned above.

[0175] If, instead, no motion capture file from network 4 and / or glove 2 is available, it is possible to proceed with manual analysis as follows (figures 17A and 17B) by: Press the PL key to play or continue playing the video clip; pressing the PL key again will pause the video clip. Press, on a KYP keypad, the character or characters corresponding to the ergonomic item being analyzed at the moments in the video in which the item is present; in the example in the figure, the representative characters are “A”, “S”, “K”, and “L”; After all the necessary ergonomic items have been gathered, Petition 870250078613, dated 03 / 09 / 2025, page 46 / 101 36 / 40 Display and verify again, if necessary, the capture that has just been completed, and then proceed to the output stage.

[0176] Using a mode selection menu, it is then possible to switch to OUTPUT mode, thus displaying the results for the entered data.

[0177] An example of a screen containing results is illustrated in Figure 18: a graph is displayed, showing the presence and temporal distribution of items in the analyzed task. In particular, in the point case, all items from points i) to iv) and vi) above are displayed, that is, i) Arm almost at shoulder height; ii) Hand above the head; iii) Extreme pronation and supination of the elbow; iv) Extreme flexion-extension and deviation of the wrist; vi) Static technical actions: maintaining an object in a static grip for at least 4 seconds.

[0178] In the case where glove 2 data were also collected, the item from point v) above would be added to the items in question, that is, recognition of the various types of holding acts and incongruent holding acts.

[0179] By consulting the graph for a point where an ergonomic item is present (for example, by clicking the point with the mouse), it is possible to display the corresponding instant in the video clip where the item occurs. In the case of capturing motion capture data by network 4, the corresponding posture of the equivalent simulation is reproduced (figure 18B).

[0180] After the data collection is completed, it is also possible to verify the calculated times again using the OCRA checklist method, that is, the times associated with the occurrence and extension of the various items.

[0181] The data can also be exported in a highly compatible format for external analysis, for example, in a Microsoft® Excel® spreadsheet. OCRA Index Method

[0182] A second ergonomic method is the OCRA index method, based Petition 870250078613, dated 03 / 09 / 2025, page 47 / 101 37 / 40 in ISO 11228-3 standards, which may allow additional body movements to be evaluated, for example: rotation of some joints through an angle that exceeds predetermined limits. Through this analysis, it is possible to obtain an assessment of a risk index for the worker.

[0183] The data collection and ergonomic item assessment methods are the same as those already described. MURI and EM-MURI methods

[0184] An additional ergonomic method is the MURI method, whereby it is possible to divide the observed time interval into several study periods in order to obtain a division of the actions performed by a worker, for example, analyzed automatically and simultaneously. At the end of the evaluation, it is possible to obtain a number of evaluations of the actions performed by a worker, instead of a single evaluation of the ergonomic analysis (i.e., based on a single ergonomic risk index).

[0185] In combination with other ergonomic methods, the EM-MURI ergonomic method can also be used, which consists of an extended version of the MURI analysis (i.e., with more parameters evaluated). The purpose of this method is to allow a quick assessment of ergonomic risks and to overestimate certain ergonomic risks in order to highlight actions in which it may be interesting to conduct a subsequent analysis.

[0186] After reading the video file corresponding to the video captured by device 6 and, if available, also the motion capture file captured with network 4, the various periods into which the task to be analyzed should be divided are defined.

[0187] When the video clip is running, the PL (play / pause) key can be used to pause the video clip at the final instant of the study period that is part of the set of periods into which the task should be divided, and the study period is identified by pressing a dedicated key; this operation is repeated until all study periods have been considered.

[0188] The start and end times of the selected period are recorded. Petition 870250078613, dated 03 / 09 / 2025, page 48 / 101 38 / 40 automatically by the software that calculates their duration, starting from the data (start / end) entered manually.

[0189] In cases where data from network 4 is available, ergonomic items that do not refer to loads to be moved will be automatically analyzed by the software. The reason for this is that in the MURI method, one piece of data that cannot be collected using the described sensors is the load (weight) to be moved or carried, so this data must be entered manually. The remaining data needed for evaluation can be automatically obtained from the sensors.

[0190] In the case where only the video clip is available, after all periods identifying the task to be analyzed have been entered, it is necessary to select the ergonomic item to be analyzed. The video clip is then restarted, and the period corresponding to that item is automatically highlighted and selected.

[0191] Although a period is highlighted, the operator performing the analysis must enter a vote that he or she intends to assign to the corresponding ergonomic item in the period in question. The voting system provides the following scale: 1 acceptable; 2 investigate; 3 unacceptable; these votes may possibly be associated with colors (e.g., green = acceptable; yellow = investigate; red = unacceptable).

[0192] In output mode, a graph is displayed summarizing the MURI data divided into periods / operations that were analyzed. An example of this graph is shown in Figure 19. TiCon method and software

[0193] Yet another ergonomic method corresponds to the EAWS method and corresponding TiCon software for ergo-characterization, that is, for the activity of defining the ergonomic parameters that distinguish from an ergonomic point of view an activity (described in the time capture or operation) based on the workstation to which the activity was assigned (adjustment).

[0194] The analysis using the EAWS method is performed with the TiCon software, which can be advantageously fed with data extracted from system 1 since it allows, for example, evaluation of: - the type of act of holding with one hand; Petition 870250078613, dated 03 / 09 / 2025, page 49 / 101 39 / 40 - the angles at which the wrist is rotated; and - a height at which an operation is performed, and a distance, direction, and force with which the body performs an operation. Additional ergonomic methods that benefit from the analysis conducted using system 1

[0195] RULA Ergonomic Method (Rapid Upper Limb Assessment): This method is used to assess musculoskeletal risks due to incongruent postures maintained in a continuous or repetitive manner during work activity.

[0196] NIOSH (National Institute for Occupational Safety and Health) ergonomic method: this method can be used to assess the risk involved in lifting loads.

[0197] Snook & Ciriello ergonomic method: this method is used to assess a risk correlated to flat transport, pulling or pushing a load.

[0198] Sue Rodgers Ergonomic Method: This method is used for muscle fatigue analysis, particularly in Brazilian industry.

[0199] Regardless of the ergonomic method chosen for ergonomic analysis using system 1, the general steps listed below for the method for ergonomic analysis of a worker using system 1 can thus be identified.

[0200] These steps include: - capturing a sequence of images of a worker's task by means of the image sequence generation unit (6); this may comprise, in combination or as an alternative to each other: capturing a video clip of the operator at the workstation; and / or generating a video clip through the interaction of a reference operator (the user U) with a virtual workstation (e.g., the CAD model of the workstation) in a virtual reality system, in particular an immersive virtual reality system; - to capture an initial set of postural data from the sensorized glove 2; this corresponds to the steps and method already described previously regarding the determination of posture and types of hand gripping acts using glove 2; Petition 870250078613, dated 03 / 09 / 2025, page 50 / 101 40 / 40 - to capture a second set of postural data from the wearable sensor network 4; and - Determine one or more ergonomic indicators based on the first dataset, the second dataset, and the sequence of images.

[0201] The acquisition of a second set of postural data from the wearable sensor network 4 comprises capturing the trajectories, postures and rotation angles of each segment element and junction element of the reference scheme during task execution.

[0202] By means of the above method and using system 1 it is consequently generally possible to identify aspects that are critical from an ergonomic point of view, and / or provide an ergonomic assessment in accordance with company and international standards, and / or redesign a workstation or sequence of operations to reduce the ergonomic risk to which the operator is exposed.

[0203] Obviously, the implementation details and embodiments may vary widely from what has been described and illustrated herein, without thereby departing from the scope of protection of the present invention, as defined by the appended claims. Petition 870250078613, dated 03 / 09 / 2025, page 51 / 101

Claims

1 / 4 CLAIMS 1. Sensorized glove (2), in particular designed for ergonomic analysis, characterized in that it includes: - an inner glove (10) comprising a plurality of extensometer sensors (EXT) configured to detect relative movements between parts of a worker's hand;and - an outer glove (12), comprising a plurality of pressure sensors (PS) distributed over a palmar surface and configured to detect the pressure exerted on corresponding areas of the palmar surface, wherein each pressure sensor (PS) is part of a sensor network (22) having a layout such as to allow application of the individual sensors (PS) to predetermined areas of the hand, wherein the sensor network (22) includes a plurality of electrical connections, connected to which are the pressure sensors (PS) and wherein all electrical connections lead to an interface unit (24) configured to allow data exchange with a processing unit (8) configured to receive data from the sensorized glove (2), wherein the inner glove includes a dorsal surface and a palmar surface and comprises: - a first plurality of pockets (14) inside the inner glove (10), housed within which are corresponding linear extensometers (EXT);and - a second plurality of pockets (16) on the outer surface of the inner glove (10), which are substantially U-shaped and house corresponding linear extensometers configured to detect relative movements of parts of the hand in a palmar plane, and wherein the outer glove (12) includes a dorsal surface and a palmar surface and comprises a third plurality of pockets (20) on the palmar surface configured to house a corresponding pressure sensor (PS).

2. Sensorized glove (2), according to claim 1, characterized in that the palmar surface of the outer glove comprises a layer of anti-slip material (18) in which the third plurality of pockets (20) is provided.

3. Sensorized glove (2), according to claim 2, characterized Petition 870250078613, dated 03 / 09 / 2025, page 52 / 101 2 / 4 by the fact that each pressure sensor (PS) is part of a network of sensors (22) having a layout such as to allow application of the individual sensors (PS) in predefined areas of the hand.

4. Sensorized glove (2), according to claim 2 or 3, characterized in that it includes a layer of material with a low coefficient of friction, preferably polyamide, extending over the palmar surface of the outer glove (12) between the outer glove itself and the anti-slip material layer (18), the layer of material with a low coefficient of friction providing a wall for each pocket (20) of the third plurality of pockets (20).

5. Sensorized glove (2), according to any one of claims 1 to 4, characterized in that it includes an inertial sensor (28) set on the dorsal surface of the outer glove (12).

6. Sensorized glove (2), according to claim 1, characterized in that the plurality of extensometer (EXT) sensors and the plurality of pressure (PS) sensors have at least partially overlapping areas.

7. Sensorized glove (2), according to claim 3, characterized in that it comprises a plurality of strips (26), preferably made of self-adhesive material, preferably Velcro®, the strips (26) being arranged on the fingers of the outer glove (12) and being configured to retain the electrical connections of the sensor network (22), routing them along the fingers of the outer glove (12).

8. Sensorized glove (2), according to claim 7, characterized in that the pockets of the third plurality include an opening on one side to facilitate insertion of the respective pressure sensor (PS), and in which one or more strips (26) extend circumferentially to an opening of a corresponding pocket (20) of the third plurality in such a way as to obstruct the same and reduce the possibility of accidental exit of the pressure sensor (22).

9. Method for ergonomic analysis of a worker's hand by means of a sensorized glove (2) as defined in any of claims 1 to 8, characterized in that it comprises: Petition 870250078613, dated 03 / 09 / 2025, page 53 / 101 3 / 4 - receiving data from the first sensor from the plurality of extensometer (EXT) sensors of the inner glove (10); - associating the data from the first sensor to a vector map of the hand, which includes a plurality of nodes associated with corresponding joints of the hand, and a plurality of segments that connect the nodes to each other, wherein one or more extensometer (EXT) sensors associated with a node are configured to detect a relative angular position between a pair of segments connected to the node, each data from the first sensor comprising a relative angular position between the pair of segments connected to the corresponding node;- receive data from the second sensor from the plurality of pressure sensors (PS) of the outer glove (12), each second sensor data comprising pressure information detected along an area of ​​the corresponding pressure sensor; - determine, for each pressure sensor (PS), a position of a point of application of a resultant force, which is a function of the corresponding second sensor data, with respect to the nodes of the vector map; and - determine, based on the data from the first sensor, in combination with the data from the second sensor, a type of gripping action exerted by the hand.

10. Method according to claim 9, characterized in that the determination of a type of gripping act performed by the hand includes: - defining a pressure map, comprising a plurality of map areas associated with corresponding pressure sensors (PS) of the outer glove (12), each map area being a set of sensitive elements (SS) having a respective area of ​​influence, each sensitive element (SS) being associated with information representing a pressure value detected in a homologous area of ​​influence on the corresponding pressure sensor (PS); - defining boundaries of involvement for each map area, wherein each boundary of involvement is representative of a different type of gripping act; - recording the information provided by the sensitive elements (SS) of each map area and comparing it with the boundaries of involvement that may be Petition 870250078613, dated 03 / 09 / 2025, p.54 / 101 4 / 4 applied to the different types of holding act; and - determine the type of holding act based on the result of the comparison between the information provided by the sensitive elements (SS) of each map area and the involvement limits that can be applied to the different types of holding act.

11. Method, according to claim 10, characterized in that it also includes performing a verification in determining the type of gripping act by comparing the determined type of gripping act with postural information determined based on data from the first sensor. Petition 870250078613, dated 03 / 09 / 2025, p. 55 / 101

Images