**1. Introduction**

Tribology is the science and technology concerning the interaction of solid surfaces in relative motion. The word tribology derives from the Greek word "tribos" that means rubbing. The topics covered by this word are various and include the study of lubricants, lubrication, friction, wear and bearings [1]. Leonardo Da Vinci was one of the first to carry out and to report studies in the field of tribology, at the end of the XVth century; he had not only performed experimental studies concerning friction but he had also developed diverse ingenious schemes for the measurement of friction. His work remained unpublished until the twentieth century when Dowson [2] presented his monumental study regarding the history of tribology. It is fascinating to note how the Da Vinci's studies on friction still remain scientifically significant today [3]. An example of Leonardo's sketches published by Dowson is shown in Figure 1.

**Figure 1.** Sketches from Leonardo's notebooks: (**a**,**b**) from Codex Atlanticus (Biblioteca Ambrosiana, Milan, Italy; CA folio 532r c. 1506-8), and (**c**) from Codex Arundel (British Library, London, UK; Arundel folio 41r c. 1500-05) [3].

Although this research topic has been studied for centuries, new analytical, numerical, and experimental methods have continued to evolve and be developed due to the intrinsic difficulty of the examination of materials' tribological properties. Indeed, they do not only depend on the type of material and the relative properties, but also on the geometry, surface conditions and topography. In addition, the measurements are also affected by several working conditions such as the pressure distribution within the contact interface, relative speed, sliding distance, temperature and relative humidity [4–6]. Consequently, extensive experimental studies adopting the most effective and robust methodologies become absolutely necessary for a deep understanding of tribological phenomena. To overcome the measurement and testing difficulties, a large amount of testing devices has been developed in the last centuries. These devices, variously called tribometers, tribotesters or friction testers (FTs), are widely used to study the friction phenomena of completely different materials, with a particular interest towards viscoelastic ones due to their advantageous characteristics, intrinsically variable in different application working ranges, and therefore particularly suitable for vibration and noise isolation or impact and impulsive shock absorption.

Starting from a particular interest in the study and in the characterization of the viscoelastic materials, the review aims to illustrate and to discuss the experimental devices designed and developed for the study of the rubber friction, being of crucial aspect in completely different contexts, e.g.,: shoe soles, O-ring sealing, conveyor belt and automotive applications, where the vehicle dynamics, the design of the apposite control systems, and the performance- and safety-focus are largely affected by what happens at the tyre/road interface in terms of friction generation mechanisms [7–10].

The review focuses on the devices that aim to study the friction between tyres and the road. It is worth highlighting that there is no limitation on the type of rubber that can be used for the friction and wear tests to be carried out with such kinds of devices, but the most commonly exploited testing conditions simulate the typical working contact conditions in the field of tyres, which are not necessarily reflected in other fields or standards. As for the latter point, it should be noted that there are some differences between the two main standards, the International Standards Organization (ISO) and the International American Standard for Testing and Materials (ASTM). For example, although in ASTM there are several indications concerning the proper selection of a method to measure the friction properties of a generic material, there are no any particular specifications for the determination of the rubber friction properties [11]. In ASTM G115-10, the "Standard Guide to Measure and Report Friction Coefficients" is also included in the ASTM Friction Test Standards so that users can choose which method may be most suitable for a particular application [12]. The methods for the determination of rubber friction are instead described in the ISO 15113 standards. This international standard refers to a linear movement and, unlike the previous ones, does not describe in details the test apparatus, but rather only provides a guide on the experimental arrangement, procedures and on the parameters to be taken into consideration to perform a robust measurement pipeline. Furthermore, the ISO

standard gives indications about the normal loads and speeds to be used within testing, additionally providing procedures for the preliminary preparation of sliding surfaces under analysis [13].

In the majority of theoretical studies on the frictional properties of materials, friction is represented using the friction model developed by Amontons and Coulomb, who claim that the frictional force is proportional to the normal force or load. However, as demonstrated initially by Bowden and Tabor [14,15] and by other authors later, Coulomb friction models are not fully reliable in case of viscoelastic materials like rubber [16,17]. Indeed, the main characteristics of the rubber friction, nowadays widely accepted and experienced, are dependent on normal force and sliding speed, temperature and a real contact area with highly-non-linear relations. These parameters play a key role in the study of the frictional behaviour of viscoelastic materials and can be used as a classification criterion to distinguish the devices and the studies to be conducted on rubbers and those to be performed in the tyre field. For instance, in a great amount of studies on rubber friction, tests are conducted at very low experimental speeds (lower than 1 mm/s) [18], so that the temperature effect in the contact area can be neglected [19] while tests conducted within the typical tyre working conditions should be generally performed at speeds in the order of meters per second.

Another crucial objective of a modern tribometer employable for tyre studies should consist in giving the possibility to work with specific samples of countersurface, i.e., an asphalt sample, since there could be a significant amount of reasons affecting the pavement characteristics: mix designs, plant operations, existing pavement conditions, or operations of the paver. It is quite obvious that it becomes really difficult to obtain the desired pavement characteristics in laboratory since even particular paving operations or peculiar mixture transformation during the compaction process phase may deeply modify the countersurface characteristics and therefore the viscoelastic behaviour of the tyre rubbers during the tyre-road interaction. For this reason, depending on a testing facility employed, it is always recommended to make sure to work with the pavement surface as similar as possible to the real one, extracting samples from a real tarmac surface for indoor testing or allowing to perform the analyses directly on track for outdoor testing with the aim to reproduce completely real test conditions. In case the test rig is employable for outdoor testing, it should be preferred since it allows one to not alter the geometries under study.

As already mentioned, the knowledge of the tyre friction behaviour and of the parameters affecting the phenomenon is an important topic both for academicians and industrial researchers, involving crucial aspects such as safety, performance, durability and environmental concerns [5]. To this end, a series of test benches have been developed by universities, research institutions and tyre makers. The laboratory tests, in addition to being less expensive compared to outdoor tests, offer the possibility to carry out measurements in almost completely controlled environmental conditions, allowing one to vary sliding speed, normal load, temperature, and other parameters in wide ranges. Another important advantage of testing tyre tread block elements in the laboratory lays in the fact that the investigations can be made at a very early stage of the tyre development, when an eventual change within the compound composition is still relatively cheap and can be easily performed, optimizing as a consequence the tyre tread geometry of sipes or other design features. The goal of the test benches is to carry out tests, not only hardly reproducible outdoors, but also representative and transferable to large-scale tyre testing. A previous investigation of these devices was conducted by Moldenhauer [20] in 2010 as part of his Ph.D. thesis. The work done by Moldenhauer has been deeply analysed, and, in the authors' opinion, further enrichment is required in order to provide a more complete analysis of the devices developed in recent years. Furthermore, Moldenhauer's study is limited to the description from the constructive point of view, while this review aims to analyse also the experimental outputs obtained, with the aid of published research references.

The paper is organized as follows: firstly, an insight on the theoretical aspects of the rubber friction is reported to provide the reader with a panorama of the main approaches and to point out the most critical aspects; then an overview of the experimental devices developed in the last fifteen years is illustrated. Such an overview cannot be complete, since there is a large number of devices used in universities' research departments or in the tyre industry, not always accessible, but it still represents a quite complete report of the current techniques and methodologies, with particular reference to dry contact conditions. Tribometers can be classify based on the contact mechanism/geometry (area, line or point contact) of the tested material with the counter-face [21]; the type of motion of the moving part [22] (linear, rotary/rolling, reciprocating, or a combination); the element motion-actuated. Referring to the type of motion, Sections 3–5 are respectively dedicated to Rolling FTs, Linear FTs and Other Types, which includes the devices that are not attributable to the other two categories.
