GB/T 26764-2024Multifunctional high-speed highway condition monitor (English PDF)

2 Normative references

The documents cited are GB 1589 (outer dimensions, axle load and mass limits of vehicles, trailers and road trains), GB/T 4208-2017 (degrees of protection provided by enclosures, IP code), GB/T 5226.1 (electrical safety of machinery, general requirements), GB/T 9969 (instructions for use of industrial products), GB/T 14267 (photoelectric distance meters), GB/T 16311 (quality requirements and test methods for road traffic markings), GB/T 18314-2009 (specifications for global positioning system surveys), GB/T 19953 (measurement of the resolution of digital cameras), GB 23254 (retro-reflective body markings for trucks and trailers) and QC/T 413 (basic technical requirements for automotive electrical equipment).

3 Terms and definitions

3.1 Multifunctional high-speed highway condition monitor: integrated intelligent inspection device able to detect automatically two or more items of road condition information among pavement distress, roughness, rut, texture depth, wearing, bumping, geometric data, the internal condition of the road structure, the coefficient of retro-reflected luminance, the forward road image and the geographical position.

3.2 Pavement distress: collective name for cracks, potholes and similar defects that reduce the technical condition or the service performance of the pavement.

3.3 Roughness: vertical deviation of the pavement surface from an ideal plane that causes a vehicle to bounce. Note: expressed by the International Roughness Index.

3.4 Rut: longitudinal band-shaped groove formed along the wheel path after the pavement has deformed, worn and subsided under repeated traffic. Note: the size of a rut is measured by the maximum groove depth of the pavement cross section.

3.5 Texture depth: depth of the open pores of the pavement surface within a specified area. Note: also called macro-texture depth; depending on the test area and the calculation model the abbreviations TD, SMTD and MPD are used.

3.6 Bumping: sudden bouncing of a vehicle caused by an abnormal bulge, subsidence or similar damage of the pavement. Note: expressed by the degree of bumping.

3.7 Wearing: state of wear of the pavement surface texture. Note: expressed by the pavement wearing rate.

3.8 Geometric data: spatial geometric shape and dimensions of the road. Note: expressed by cross slope, longitudinal slope, horizontal curve radius and vertical curve radius.

3.9 Coefficient of retro-reflected luminance: ratio of the luminance in the direction of observation to the normal illuminance on a plane perpendicular to the direction of the incident light.

4 Composition and functions

4.1 The monitor is made up of the carrier vehicle, the mileage measuring device, the functional detection systems and the data processing system. The functional detection systems are the geographical position acquisition system, the pavement distress detection system, the roughness detection system, the rut detection system, the texture depth and wearing detection system, the bumping detection system, the ground penetrating radar detection system, the geometric data detection system, the retro-reflected luminance coefficient detection system and the forward image acquisition system.

4.2 The monitor has to work with the road open to traffic. Mileage measurement supplies mileage information to the data processing system and a common drive signal to all the other systems. Geographical position acquisition uses a satellite-based geodetic positioning device and links the position to the mileage by software.

4.2 Pavement distress detection uses a planar image acquisition device or a three-dimensional pavement data acquisition device and yields, after manual or automatic recognition, statistics classified by distress type. Roughness detection uses laser distance sensors and accelerometers to record the change of longitudinal profile slope and computes roughness through a roughness model. Rut detection uses laser distance sensors to record the shape of the cross section and computes rut depth through a rut model.

4.2 Texture depth and wearing detection uses laser distance sensors to record the pavement texture and computes texture depth and pavement wearing data. Bumping detection combines laser distance sensors with an inertial reference device to record relative elevations of the longitudinal profile. Ground penetrating radar detection uses two-dimensional or three-dimensional radar to obtain echo information on the road structure and, after processing and imaging, the thickness of the structural layers and information on defects under the pavement.

4.2 Geometric data detection uses gyroscopes and angular accelerometers to record transverse and longitudinal angular change and the change of curvature of the horizontal and vertical alignment. Retro-reflected luminance coefficient detection uses a special light source with emitting and receiving devices to measure the retro-reflected luminance of road markings. Forward image acquisition records the landscape ahead of the vehicle and yields information on the road environment and the traffic safety facilities.

4.2 The data processing system covers acquisition and processing. The acquisition software drives all systems from the same source, stores the data and unifies the position of the multi-source data; its main interface controls and monitors the whole system and allows the detection route, start chainage, direction, lane, sampling interval, storage location and display mode to be set, with real-time calibration and checking of the chainage during detection. The processing software computes each indicator separately and produces single-indicator or multi-indicator statistics and assessments.

5 Technical requirements

5.1.1 The vehicle has to be clean and complete, with all parts sound and firmly connected. External attachments carry retro-reflective marks, and the rear of the carrier vehicle carries a permanent warning legend such as road inspection, or a variable-message electronic screen, together with warning lamps; the legend must be legible and unobstructed and its total outline area not smaller than 60 % of the projected rear area of the carrier vehicle. The outline dimensions of the carrier vehicle in working condition comply with GB 1589, the vehicle appears in the national vehicle production catalogue, and the ground penetrating radar is connected to the carrier vehicle as an integrated structure.

5.1.1 Inside the carrier vehicle an antistatic insulating floor is laid and an antistatic device is fitted outside. A stable power supply system with safety protection is provided, together with an interface for an external supply, and the on-board supply sustains continuous work with adequate power quality. Other electrical equipment complies with GB/T 5226.1 and QC/T 413. External attached devices have a degree of protection not lower than IP55 of GB/T 4208-2017.

5.1.2 and 5.1.3 Mileage is measured by an incremental photoelectric encoder or a similar device that also supplies the common drive signal. Geographical position is acquired by a satellite positioning system such as BeiDou, linked automatically to the chainage, and recorded and stored at 5 m intervals.

5.1.4 Pavement distress detection may use area-array imaging, line-scan imaging, laser digital imaging or three-dimensional data imaging, with auxiliary lighting as needed. It has to cover both asphalt and cement concrete pavements, work continuously in the longitudinal direction, and deliver orthogonal images with accurate position information, clear texture and even brightness suitable for automatic recognition. The software displays and stores images or three-dimensional data in real time, allows manual assistance, review and correction when automatic recognition is used, stores the annotated result, and provides statistics over 10 m, 100 m and 1 000 m intervals; the 10 m damage ratio is the sum of the damaged areas multiplied by their weights or conversion factors divided by the inspected pavement area, and the 100 m and 1 000 m ratios are the arithmetic means of the 10 m ratios they contain.

5.1.5 Roughness detection uses laser profile measurement referred to an inertial reference system, covers asphalt and cement concrete pavements, measures on the centre line of one or both wheel paths, and uses accelerometers with a nominal range of at least +/- 2 g. The software computes and stores the IRI of 10 m intervals according to Annex A, flags invalid data when conditions are abnormal, and aggregates to 100 m and 1 000 m as arithmetic means.

5.1.6 Rut detection uses multi-point laser on a common beam or line structured light, with a transverse sampling interval not greater than 300 mm and a longitudinal sampling interval not greater than 200 mm; for the multi-point laser beam type the nominal measuring range of the laser distance sensor is at least 200 mm. The software computes the left, right and section rut depth of every cross section according to Annex B, the section value being the greater of the two sides, flags incomplete or abnormal sections, and aggregates over 10 m, 100 m and 1 000 m as arithmetic means.

5.1.7 Texture depth and wearing detection uses laser profile measurement with a longitudinal sampling interval not greater than 2 mm; in two-line detection the sensors sit on the centre lines of the left and right wheel paths, in three-line detection on the wheel paths and the lane centre line. The software computes the texture depth of 10 m intervals according to Annex C, aggregates to 100 m and 1 000 m, and derives the pavement wearing rate from the 10 m texture depth through the wearing rate model.

5.1.8 Bumping detection uses a single laser or a laser group measuring the longitudinal profile, covers asphalt and cement concrete pavements and measures on the centre lines of the left and right wheel paths with a sampling interval not greater than 100 mm. Its software removes bridge expansion joints and other abnormal data and the influence of the longitudinal alignment before computing the number and degree of bumps per 10 m according to Annex D.

5.1.9 Ground penetrating radar detection uses two-dimensional or three-dimensional radar; a multi-channel two-dimensional radar has at least 8 channels and a three-dimensional radar at least 16. The nominal main frequency of the antenna is at least 1 GHz for structural layer thickness and from 100 MHz to 1 GHz for internal defects. The longitudinal sampling interval is not greater than 200 mm for layer thickness and not greater than 50 mm for internal defects. The software stores the raw scan of every antenna channel, links it to chainage and geographical position, produces interval statistics including arithmetic mean, standard deviation and representative value, and generates defect cards and statistical reports automatically.

5.1.10 to 5.1.12 Geometric data detection uses accelerometers and an inertial navigation system, covers cross slope, longitudinal slope and the curvature of horizontal and vertical curves, samples at intervals not greater than 0.5 m and outputs samples at intervals not greater than 1 m. The retro-reflected luminance coefficient is measured with 30 m geometry, that is a nominal entrance angle of 88.76° and a nominal observation angle of 1.05°, on one or both sides, with spectral matching for white and yellow markings, immunity to ambient light in daylight and an output interval not greater than 1 m. Forward image acquisition uses area-array imaging with a sampling interval not greater than 10 m, images linked to chainage and position, and results stored in a common format such as JPEG.

5.2 The working environment is an ambient temperature from -10 °C to 50 °C, a relative humidity not greater than 90 %, an altitude not above 5 500 m, a pavement free of standing water, snow, ice and contamination, and a wind force below force 6.

5.3 Table 1 fixes the performance limits. Mileage measurement error is +/- 0.05 %. For the geographical position acquisition system, with good satellite coverage the proportion of points whose plane deviation exceeds 2 m is not greater than 5 %. For pavement distress: transverse detection width not less than 70 % of the lane width, minimum resolvable crack width 1 mm, distress area measurement error +/- 10 %. For roughness: laser distance sensor conforming to GB/T 14267, indication error +/- 1 mm, repeatability <= 5 %, speed influence error +/- 5 %, correlation coefficient >= 0.99. For rut: sensor conforming to GB/T 14267, indication error +/- 1 mm, transverse detection width >= 3.5 m, repeatability <= 5 %, speed influence error +/- 5 %, correlation coefficient >= 0.99. For texture depth and wearing: indication error +/- 0.5 mm, repeatability <= 5 %, speed influence error +/- 5 %, correlation coefficient >= 0.97. For bumping: indication error +/- 1 mm, repeatability <= 5 %, speed influence error +/- 5 %.

5.3 Table 1 continues with ground penetrating radar: for thicknesses of 10 cm and below the indication error is +/- 3 mm for an air-coupled antenna and +/- 10 mm for a ground-coupled antenna, and above 10 cm it becomes +/- 3 % and +/- 10 % respectively; the antenna coverage width is not less than 40 % of the lane width and the detection rate of internal defects such as voids and debonding is at least 90 %. For geometric data: indication error +/- 0.5 % on slope and +/- 1 m on a 50 m curve radius, repeatability <= 5 %, speed influence error +/- 5 %. For the retro-reflected luminance coefficient: indication error +/- 7 %, repeatability <= 5 %, speed influence error +/- 5 %, ambient light influence error +/- 5 %. The forward image acquisition device has a video resolution of at least 5 million pixels. A footnote adds that where conditions allow, roughness tests on curves, under emergency braking or rapid acceleration and at ultra-low speed may be added following Annexes E, F and G, with a relative error between the measured and the reference value not greater than 15 %.

6 Test methods

6.1 and 6.2 Tests are run at an ambient temperature from 10 °C to 30 °C and a relative humidity not greater than 85 %, on a flat straight section free of standing water, contamination and intersections and with little change of slope. The listed apparatus includes steel tapes of 0 m to 5 m and 0 m to 50 m and a steel rule of 0 mm to 300 mm all graduated in 1 mm, a total station of accuracy class II, a geodetic GNSS receiver meeting class E of GB/T 18314-2009, gauge blocks of 5 mm, 20 mm, 40 mm and 80 mm of accuracy class II, a test platform of 150 mm x 150 mm x 10 mm with a flatness of 0.05 mm per 150 mm and a surface roughness Ra of 0.8 µm, a level of class DSZ05, a 4 m cross-section rule graduated every 50 mm, a set of at least ten thickness specimens of the same material, and retro-reflective reference specimens 1 000 mm +/- 10 mm long and 150 mm +/- 10 mm wide in white and in yellow whose colour meets GB/T 16311.

6.2 Table 2 splits the reference specimens into three ranges of retro-reflected luminance coefficient, in millicandela per square metre per lux: for the white specimen 0.1 to 150, 151 to 450 and above 450; for the yellow specimen 0.1 to 100, 101 to 175 and above 175.

6.3 General requirements are checked by eye and by hand for the vehicle appearance, the soundness of the parts and the system integration; the retro-reflective marking, warning lamps and permanent legend are checked to GB 23254 and the vehicle catalogue entry is consulted. The external power interface, the antistatic device and the electrical protection are inspected visually, electrical safety follows GB/T 5226.1 and QC/T 413, and the protection degree of attached equipment follows GB/T 4208-2017. Component specifications are read from the product documents and mounting dimensions measured with the steel rule and tape. The software is exercised by testing 1 km of road and processing the data.

6.4 For mileage error a standard length of 500 m is set out on the test section with a steel tape or total station, marked at both ends with a straight running trace between them; the run starts when the vertical projection of the axle centre of the wheel carrying the measuring device is aligned with the start mark and stops at the end mark. Three consecutive runs are averaged and the relative error is computed from the standard length and the measured mileage.

6.5 For the static positioning deviation ratio a test point is surveyed with the geodetic receiver to class E of GB/T 18314-2009 and converted into plane rectangular coordinates as the reference. The receiving antenna of the monitor is then set on the point, or the vehicle moved so that the antenna projects onto it, and position is acquired continuously for 60 min; the deviation of each fix from the reference is computed and the percentage of fixes deviating by more than 2 m is the static positioning deviation ratio.

6.6 Pavement distress tests: for transverse detection width a white rule graduated in 1 mm is stuck across the lane over at least 5 m in the middle of a straight section at least 100 m long, the vehicle runs at 50 km/h and the width is read from the images. For the minimum resolvable crack width, a straight section at least 100 m long carries a marked crack at least 1 m long and not wider than 1 mm, and the vehicle passes at a steady 50 km/h while the image is inspected by eye. For distress area error, five 100 m sections with different degrees of asphalt pavement damage and containing at least cracking, alligator cracking, patching and potholes are measured by hand as the reference, then inspected by the monitor and processed manually or automatically, and the arithmetic mean of the five relative errors is the result.

6.7 Roughness tests: the laser distance sensor follows GB/T 14267. For indication error the vehicle stands on a hard level pavement, the test platform is levelled under the laser spot, the vertical distance to the platform is taken as the zero reference and gauge blocks of 5 mm, 20 mm, 40 mm and 80 mm are inserted in turn, the largest absolute difference between measured and standard value being the result. Repeatability uses a section at least 100 m long with IRI between 0 m/km and 3 m/km, acceleration and deceleration lengths of at least 100 m at each end, points marked every 5 m, ten runs at 50 km/h with the trace held within a circle of 0.10 m diameter, and the coefficient of variation of the ten hundred-metre values as the result. Speed influence error compares the mean of five hundred-metre values at 30 km/h with the mean of five at 70 km/h. Correlation uses four sections, the smoothest with IRI not greater than 2.0 m/km and the roughest with IRI not less than 5.0 m/km and the others spread between them, each 100 m to 200 m long with points every 5 m, three runs at 50 km/h compared with elevations levelled every 25 cm and processed through Annex A.

6.8 to 6.10 Rut tests take the transverse width from the distance between the outermost laser spots or from the effective laser line projected on a hard pavement, and follow the roughness procedures for indication error, repeatability and speed influence, with rutting evenly distributed over the section. Rut correlation uses four sections with rut depth from not more than 10 mm to not less than 40 mm, points every 10 m, three runs at 50 km/h and the cross-section rule and steel rule with the Annex B model as reference. Texture depth follows the same pattern, with four sections from not more than 0.3 mm to not less than 1.2 mm and the sand patch method as reference. Bumping tests run on a section with slight bumping, with a relative height difference between 2 cm and 5 cm after the Annex D pre-processing and a test speed of 30 km/h.

6.11 Ground penetrating radar tests: each thickness specimen is measured once along each of its four edges with the steel rule and the mean taken as its standard thickness. A metal plate at least as large as the specimens is laid flat, a specimen of intermediate thickness is placed on it for wave speed calibration and measured for 1 min, ten traces are drawn at random and the mean two-way travel time gives the wave speed. Each specimen is then measured for 1 min with the calibrated radar, ten traces are drawn at random and the mean thickness gives the indication error, expressed as a relative error above 10 cm. The transverse coverage is measured with a steel tape between the outer edges of the outermost antennas or antenna units. The detection rate of internal defects uses a real or simulated section with at least ten confirmed voids or debonding zones buried between 0.5 m and 2.0 m, whose type, number, plan position and depth are compared with the radar findings.

6.12 Geometric data tests: cross slope uses a section at least 100 m long with a uniform cross slope between 1.5 % and 2.5 %, points marked every 10 m, three runs at 50 km/h compared with level readings. Longitudinal slope uses a section at least 50 m long with a uniform grade between 4.0 % and 6.0 %, points every 5 m, three runs at 50 km/h compared with slopes computed from levelled elevations. The 50 m curve radius uses a flat area of at least 100 m x 100 m on which a curve of 50 m radius and at least 100 m length is drawn with a 50 m tape fixed at the centre, with straight acceleration and deceleration lengths of at least 50 m and three runs at 20 km/h.

6.13 and 6.14 Retro-reflected luminance coefficient tests place three reference specimens of the same colour but different coefficient ranges 1 m apart along the lane and measure them ten times at 50 km/h; the indication error is the largest relative departure of the mean from the reference value. Repeatability is the largest coefficient of variation of the same ten measurements. Speed influence compares ten measurements at 30 km/h with ten at 70 km/h against the reference. Ambient light influence compares ten measurements at 50 km/h taken when the illuminance is above 50 000 lx with ten taken when it is below 4 000 lx. Forward image resolution follows GB/T 19953.

7 Inspection rules

7.1 Inspection is divided into type inspection and delivery inspection, and Table 3 lists for every item the clause holding the requirement, the clause holding the test method and whether the item belongs to type inspection, to delivery inspection or to both.

7.1 Type inspection is required when a new product is finalised or transferred to another plant, when after regular production an important structure, material or process changes in a way that may affect performance, when production resumes after a stop of more than six months, before the first imported unit is put into use, and when the market supervision authorities ask for it.

7.2 Every sample presented for delivery inspection is inspected in full; samples for type inspection are drawn at random from products that have passed delivery inspection, three complete units of the same product model.

7.3 Type inspection passes when every indicator meets the requirements, otherwise it fails; delivery inspection is judged in the same way. Where either inspection shows a non-conforming item, the product is reworked or repaired and is judged conforming only if a further inspection meets every requirement.

8 Marking, accompanying documents and storage

8.1 A nameplate is fixed to the monitor carrying at least the product name, the product model, the name of the manufacturer, the serial number and the date of manufacture. Each functional detection system carries a mark giving the name of that system.

8.2 A valid certificate of conformity and instructions for use meeting GB/T 9969 are supplied with the product.

8.3 The product is best stored in a ventilated, dry and dust-free space free of corrosive gases or liquids.

A to G Annexes

Annex A (normative) gives the logic of a BASIC program with a step of 0.25 m for computing the International Roughness Index, printed as a numbered program listing.

Annex B (normative) sets out the rut envelope model. The rut depth of every inspected cross section is classified and computed against seven standard models; the accompanying figure identifies the actual pavement line, the left rut depth, the rut envelope curve and the right rut depth.

Annex C (normative) divides the texture profile into calculation units 0.3 m long and computes the SMTD of each unit. The symbols are the texture depth over the base calculation length, the base calculation length taken as 0.3 m, the nominal distance of each point in metres, an intermediate result without physical meaning, the profile elevation of each point in millimetres, the number of profile elevation samples within the base length rounded to the nearest odd number, and the longitudinal sampling interval in metres.

Annex D (normative) gives the bumping calculation. The bump judging height difference, in centimetres, is the range between the highest and the lowest of 101 relative elevations taken every 0.1 m over a 10 m interval, after removal of bridge expansion joints and other abnormal data and of the influence of the longitudinal alignment. Table D.1 grades bumping by that height difference: light from 2 cm up to but not including 5 cm, moderate from 5 cm up to but not including 8 cm, and severe at 8 cm and above. A bump is counted per occurrence: if one occurs within 10 m it counts as one, and the degree corresponding to the largest height difference is recorded.

Annex E (informative) tests roughness on curves, using a 200 m section that contains a curve of radius not greater than 200 m and has evenly distributed roughness, run at the highest safe cornering speed, preferably not below 30 km/h, and compared with elevations levelled every 25 cm and processed through Annex A.

Annex F (informative) tests roughness under emergency braking or rapid acceleration on a straight 200 m section with evenly distributed roughness, the vehicle decelerating from 60 km/h to below 5 km/h within 70 m, accelerating from 5 km/h to 30 km/h within 20 m and leaving the section at a steady 30 km/h, the result again compared with levelled elevations processed through Annex A.

Annex G (informative) tests roughness at ultra-low speed on a straight 320 m section with evenly distributed roughness run at a steady 10 km/h, compared in the same way with levelled elevations processed through Annex A.