Wheel-legged robot tire pressure integrated sensing system and ground contact sensing method

Abstract

The application provides a kind of wheel-foot robot tire pressure integrated sensing system and ground touch sensing method, it is related to robot control technical field, including tire pressure integrated sensing module, receiving processing module and robot motion control host computer;Tire pressure integrated sensing module includes air pressure acquisition module, main control MCU module, wireless radio frequency transceiver module and power management module;Air pressure acquisition module is used to collect the original air pressure digital quantity and original temperature digital quantity in tire closed cavity;Main control MCU module is used to execute acquisition timing scheduling, first-level sensor compensation and data frame encapsulation, generates data frame;Wireless radio frequency transceiver module is used to send data frame to receiving processing module wirelessly;Receiving processing module is used to carry out integrity check and time alignment to data frame, executes second-level dynamic drift compensation, ground touch event segmentation, contact state solution and transient form feature extraction, and outputs ground touch state, contact confidence and terrain feature vector to robot motion control host computer.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a tire pressure integrated sensing system and ground contact sensing method for a wheeled robot. Background Technology

[0002] Wheeled robots combine the efficiency of wheeled locomotion with the obstacle-crossing ability of legged robots. Their state estimation, gait switching, posture control, and terrain adaptation rely on timely feedback of wheel-foot contact states. For wheeled robots using flexible pneumatic tires as feet, the contact time, lift-off time, contact duration, load changes, and impact characteristics are all important control and perception information.

[0003] Existing robot ground contact detection typically employs microswitches, thin-film pressure sensors, strain gauge sensors, or multi-dimensional force sensors. For example, patent CN111458766A uses springs and contact switches to determine whether a robot's foot is on the ground. While such contact-based detection devices can directly output contact status or force information, they are susceptible to impact, wear, and structural fatigue during running, jumping, and movement on complex terrain, and also increase the mass of the leg's end effector. For continuously rotating pneumatic wheels, wired sensors also suffer from problems such as wire entanglement, slip ring wear, and structural complexity, and are generally unable to obtain continuous high-frequency dynamic characteristics during the ground contact process.

[0004] Patents such as CN107902002B (authorization announcement number) and WO2021088104A1 (publication number) disclose wheel-leg composite mechanisms and wheel- and foot-based motion conversion schemes. However, they mainly focus on wheel-leg structures, transmission relationships, and motion mode switching, and do not propose complete technical solutions for high-frequency acquisition, wireless transmission, ground contact determination, and terrain feature extraction of flexible pneumatic tire pressure.

[0005] Existing automotive tire pressure monitoring systems collect and transmit tire pressure data using pressure sensors, temperature sensors, and radio frequency devices within the tire. For example, patent CN110733300A discloses a remote real-time tire pressure monitoring system for vehicles, primarily used for tire condition monitoring, remote display, and anomaly alarms. However, existing automotive tire pressure monitoring systems typically output absolute pressure, temperature, and underpressure alarms, which differs from the high-frequency continuous waveforms, low-latency ground contact feedback, and closed-loop motion control required by wheeled robots.

[0006] Furthermore, during the movement of the wheeled robot, tire deformation, frictional heat generation, and changes in ambient temperature can cause tire pressure drift. There may also be dynamic thermal hysteresis between the gas inside the tire and the sensor chip. Sensor aging, power supply changes, and slow tire leakage can cause pressure reference drift. If a fixed temperature compensation or continuous sliding average is used to update the reference, the actual bearing pressure may be written into the reference during the support phase, causing reference contamination.

[0007] Different initial tire pressures, robot loads, movement speeds, and ground conditions can lead to significant variations in the amplitude and rate of change of contact pressure. When using a fixed threshold, a threshold that is too low can easily cause false triggers, while a threshold that is too high can easily miss soft ground or low-load contact. Directly using the raw pressure peak value or unnormalized waveform for terrain identification is also easily affected by the amplitude of a single impact and the duration of contact, reducing the comparability of features under different operating conditions.

[0008] Therefore, there is a need for a wheeled robot tire pressure integrated sensing system and ground contact sensing method suitable for wheeled robots, in order to achieve high-frequency acquisition of air pressure and temperature, low-latency wireless transmission, dynamic thermal hysteresis compensation, reference tracking with unloading state gating, noise-adaptive ground contact determination, and normalized terrain transient feature extraction, thereby improving the stability of ground contact sensing under different temperature, load, tire pressure, speed and terrain conditions. Summary of the Invention

[0009] To address the aforementioned technical problems, this invention provides an integrated tire pressure sensing system for wheeled robots.

[0010] The technical solutions provided by the embodiments of the present invention are as follows: The first aspect of the present invention provides a tire pressure integrated sensing system for a wheeled robot, comprising a tire pressure integrated sensing module connected to a sealed cavity air passage of the robot's flexible pneumatic tire, a receiving and processing module disposed on the robot body, and a robot motion control host computer communicatively connected to the receiving and processing module. The tire pressure integrated sensing module includes a pressure acquisition module, a main control MCU module, a wireless radio frequency transceiver module, and a power management module. The air pressure acquisition module is used to acquire the original digital air pressure and the original digital temperature in the sealed cavity of the tire, and send the original digital air pressure and the original digital temperature to the main control MCU module through the data bus. The main control MCU module is used to perform acquisition timing scheduling, first-level sensor compensation and data frame encapsulation, generate data frames, and send the data frames to the wireless radio frequency transceiver module. The wireless radio frequency transceiver module is used to wirelessly transmit the data frame to the receiving and processing module. The receiving and processing module is used to perform integrity verification and time alignment on the data frame, perform secondary dynamic drift compensation, ground contact event segmentation, contact state calculation and transient morphological feature extraction, and output ground contact state, contact confidence and terrain feature vector to the robot motion control host computer. The power management module is electrically connected to and supplies power to the air pressure acquisition module, the main control MCU module, and the wireless radio frequency transceiver module.

[0011] Optionally, the air pressure acquisition module includes a pressure sensing unit, a temperature detection unit, an analog-to-digital conversion unit, and a calibration parameter storage unit. The air pressure acquisition module communicates with the main control MCU module via an SPI bus or an I2C bus and outputs the raw air pressure digital value at a pressure sampling frequency higher than the temperature sampling frequency.

[0012] Optionally, the main control MCU module includes a hardware timer triggering unit, a direct memory access unit, an acquisition scheduling unit, and a transmission queue; the acquisition scheduling unit adopts a multi-rate scheduling method of interleaved pressure high-frequency sampling and temperature low-frequency sampling, and decouples the air pressure acquisition task, compensation calculation task, and wireless transmission task so that any wireless transmission cycle does not block the next pressure sampling cycle.

[0013] Optionally, the primary sensor compensation includes: performing physical quantity calculation on the original digital air pressure based on the calibration coefficients in the calibration parameter storage unit; correcting the zero point, sensitivity, and nonlinearity errors based on temperature data; and estimating the dynamic thermal hysteresis component between the pressure sensing unit and the tire cavity based on the current temperature, temperature change rate, and historical temperature status to obtain primary compensated air pressure data.

[0014] Optionally, the data frame includes at least one of node identifier, time information, sequence number, compensated air pressure data, temperature data, sampling time, sampling interval, cyclic redundancy check code, power supply status, and sensor diagnostic status.

[0015] Optionally, the wireless radio frequency transceiver module reads the transmission queue using interrupt or direct memory access methods, and completes wireless transmission according to at least one strategy among node time slots, staggered backoff, or event priority.

[0016] Optionally, the tire pressure integrated sensing module is installed inside the tire, on the rim, at the tire valve, or in a location connected to the tire's sealed cavity via an air guide structure; the air pressure acquisition module is connected to the sealed cavity formed by the rim and the inflatable tire via an air inlet or air guide channel, and the air inlet or air guide channel is provided with a breathable and waterproof structure; the printed circuit board carrying the tire pressure integrated sensing module is provided with a digital ground area and a radio frequency ground area, and a copper-free clearance area is provided in the antenna radiation area of ​​the wireless radio frequency transceiver module.

[0017] Optionally, the receiving and processing module includes a dynamic thermal hysteresis compensation unit, an unloading state discrimination unit, a robust reference tracking unit, a noise scale estimation unit, a contact state calculation unit, and a morphological feature extraction unit; the robust reference tracking unit updates the dynamic reference only when the unloading state discrimination unit outputs a valid update gating signal, so as to avoid the support phase pressure pulsation being written into the dynamic reference.

[0018] A second aspect of this invention provides a ground contact sensing method applied to the aforementioned wheeled robot tire pressure integrated sensing system, comprising: S1, a pressure acquisition module acquiring raw digital air pressure and raw digital temperature; S2, a main control MCU module performing first-level temperature, nonlinearity, and dynamic thermal hysteresis compensation on the raw digital air pressure to obtain first-level compensated air pressure data; S3, the main control MCU module encapsulating the first-level compensated air pressure data with node identifiers, time information, and sequence numbers into a data frame, and transmitting it to a receiving and processing module via a wireless radio frequency transceiver module; S4, the receiving and processing module performing integrity verification and packet loss marking on the data frame. S5, the receiving and processing module updates the robust dynamic reference and noise scale according to the unloading state gating, and subtracts the dynamic thermal hysteresis component and dynamic reference from the first-level compensated air pressure data to obtain the second-level compensated air pressure data; S6, the contact probability is calculated based on the second-level compensated air pressure data, air pressure change rate and noise scale, and the ground contact state is determined by adaptive dual thresholds for ground contact sensing; S7, the amplitude and time of the second-level compensated air pressure waveform of a single ground contact event are normalized, the terrain feature vector is extracted, and the ground contact sensing result containing the ground contact state, contact confidence and terrain feature vector is output to the robot motion control host computer.

[0019] Optionally, the dynamic thermal hysteresis compensation in step S2 adopts a recursive thermal state model. The input of the recursive thermal state model includes at least the current temperature, the first-order rate of change of temperature, and the thermal state at the previous moment. The output is the slowly changing pressure component caused by the temperature change. The slowly changing pressure component is separated from the fast pressure component caused by ground contact at different time scales. Optionally, the unloading state gating conditions in step S5 include at least: the absolute value of the rate of change of air pressure is lower than the rate of change threshold, the variance of the sliding window is lower than the variance threshold, and the contact probability is lower than the probability threshold; if and only if the unloading state gating conditions are continuously satisfied for a preset duration, the dynamic benchmark is updated using the median of the sliding window, and the noise scale is updated using the median absolute deviation.

[0020] Optionally, the ground contact trigger threshold and the ground release threshold in step S6 are adaptively determined by at least two of the noise scale, tire static pressure, temperature state and movement speed; the contact probability is calculated based on at least three of the secondary compensated pressure data, first-order pressure change rate, second-order pressure change rate, window energy and duration, and the ground contact state is output through hysteresis logic where the trigger threshold is higher than the release threshold. Optionally, step S7 includes: normalizing the waveform time axis with the start and end times of the ground contact event, normalizing the waveform amplitude with peak value, energy or steady-state increment, and extracting at least four items from the normalized waveform, including rise time, peak position, loading slope, unloading slope, area ratio, attenuation ratio, kurtosis and preset frequency band energy, to form a terrain feature vector that weakens the influence of single-step stamping force and movement speed.

[0021] Optionally, the receiving and processing module performs unified time axis resampling on the data of multiple wheel feet based on the time information in the data frames of each tire pressure integrated sensing module, constructs a contact pattern vector composed of the contact probabilities of each wheel foot, and sends the contact pattern vector and the terrain feature vector together to the robot motion control host computer for state estimation, gait planning or control strategy switching.

[0022] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: (1) In this embodiment of the invention, hardware timed triggering, multi-rate interleaved sampling, direct memory access and asynchronous transmission queue are used to enable wireless transmission to not block pressure sampling, and to establish a verifiable and aligned data stream using node identifier, time information and sequence number.

[0023] (2) In this embodiment of the invention, the combination of node-end factory calibration and temperature nonlinear correction with receiver-end dynamic thermal hysteresis estimation can better distinguish between slow heating effects and fast grounding pulses compared to static correction based solely on the current temperature.

[0024] (3) In this embodiment of the invention, the dynamic benchmark is updated only when the unloading state with low rate of change, low variance and low contact probability is continuously satisfied, and the influence of impact outliers, slow leakage and occasional packet loss is suppressed by the median and median absolute deviation.

[0025] (4) In this embodiment of the invention, the dual thresholds are adaptively determined based on real-time noise scale, static tire pressure, temperature state and movement speed, and the contact probability and contact confidence are output at the same time, so as to reduce the false triggering and hysteresis caused by fixed thresholds under different loads, speeds and terrains.

[0026] (5) In this embodiment of the invention, by segmenting the ground contact event, normalizing the amplitude and normalizing the time, the relative shape of the waveform is preserved, the absolute difference between the pressure peak and the contact duration is weakened, and the terrain features are made to focus more on the dynamic response of the tire-ground interaction.

[0027] (6) In this embodiment of the invention, the tire pressure integrated sensing module is fixed to the wheel hub and transmits data through a wireless link, eliminating the need to arrange signal lines or slip rings across the rotating joint; the digital ground and radio frequency ground partitioning, antenna clearance and breathable waterproof structure improve the communication and measurement stability in harsh environments. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a schematic diagram illustrating the internal composition, connection relationships, and data flow of a tire pressure integrated sensing module provided in an embodiment of the present invention.

[0030] Figure 2 This is a schematic diagram of the system data link of a tire pressure integrated sensing system for a wheeled robot provided in an embodiment of the present invention.

[0031] Figure 3 The flowchart illustrates a method for a main control MCU module to perform non-blocking multi-rate acquisition, first-level compensation, and asynchronous transmission, as provided in an embodiment of the present invention.

[0032] Figure 4 This is a schematic diagram illustrating an installation method of a tire pressure integrated sensing module disposed within the sealed cavity of a flexible pneumatic tire, as provided in an embodiment of the present invention.

[0033] Figure 5 This is a flowchart of a two-stage compensation, unloading gating benchmark tracking, probabilistic ground contact determination, and normalized terrain feature extraction method provided in an embodiment of the present invention.

[0034] Figure reference numerals: 1-Tire pressure integrated sensor module; 101-Air pressure acquisition module; 102-Main control MCU module; 103-Wireless radio frequency transceiver module; 104-Power management module; 2-Receiver processing module; 3-Robot motion control host computer.

[0035] As shown in the figure, specific structures and devices are marked in the figure to clearly illustrate the structure of the embodiments of the present invention. However, this is only for illustrative purposes and is not intended to limit the present invention to this specific structure, device and environment. Those skilled in the art can adjust or modify these devices and environments according to specific needs. Detailed Implementation

[0036] The technical solutions of the present invention will now be described with reference to the accompanying drawings. It should also be noted that, to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also use other alternative methods to implement the invention; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0037] like Figures 1 to 5 As shown, an embodiment of the present invention provides a tire pressure integrated sensing system for a wheeled robot, including a tire pressure integrated sensing module 1 connected to the air passage of the robot's flexible pneumatic tire, a receiving and processing module 2 disposed on the robot body, and a robot motion control host computer 3 communicatively connected to the receiving and processing module 2.

[0038] The tire pressure integrated sensing module 1 includes a tire pressure acquisition module 101, a main control MCU module 102, a wireless radio frequency transceiver module 103, and a power management module 104.

[0039] The air pressure acquisition module 101 is used to acquire the original digital air pressure and original digital temperature in the sealed cavity of the tire, and send the original digital air pressure and original digital temperature to the main control MCU module 102 through the data bus.

[0040] The main control MCU module 102 is used to perform acquisition timing scheduling, first-level sensor compensation and data frame encapsulation, generate data frames, and send the data frames to the wireless radio frequency transceiver module 103.

[0041] The wireless radio frequency transceiver module 103 is used to wirelessly transmit data frames to the receiving and processing module 2.

[0042] The receiving and processing module 2 performs integrity verification and time alignment on the data frames, executes secondary dynamic drift compensation, ground contact event segmentation, contact state calculation, and transient morphological feature extraction, and outputs the ground contact state, contact confidence, and terrain feature vector to the robot motion control host computer 3. After receiving the ground contact state, contact confidence, and terrain feature vector, the robot motion control host computer 3 performs motion control on the robot.

[0043] The power management module 104 is electrically connected to and supplies power to the air pressure acquisition module 101, the main control MCU module 102 and the wireless radio frequency transceiver module 103 respectively.

[0044] Furthermore, the air pressure acquisition module 101 includes a pressure sensing unit, a temperature detection unit, an analog-to-digital conversion unit, and a calibration parameter storage unit. The air pressure acquisition module 101 communicates with the main control MCU module 102 via an SPI bus or an I2C bus, and outputs the raw air pressure digital quantity at a pressure sampling frequency higher than the temperature sampling frequency.

[0045] Furthermore, the main control MCU module 102 includes a hardware timing trigger unit, a direct memory access unit, an acquisition scheduling unit, and a transmission queue. The acquisition scheduling unit adopts a multi-rate scheduling method of interleaved sampling of high-frequency pressure and low-frequency temperature, and decouples the air pressure acquisition task, compensation calculation task, and wireless transmission task so that any wireless transmission cycle does not block the next pressure sampling cycle.

[0046] Furthermore, the first-level sensor compensation includes: performing physical quantity calculations on the raw digital air pressure based on the calibration coefficients in the calibration parameter storage unit; correcting zero-point, sensitivity, and nonlinearity errors based on temperature data; and estimating the dynamic thermal hysteresis component between the pressure sensing unit and the tire cavity based on the current temperature, temperature change rate, and historical temperature conditions to obtain first-level compensated air pressure data.

[0047] Furthermore, the data frame includes at least one of the following: node identifier, time information, sequence number, compensated air pressure data, temperature data, sampling time, sampling interval, cyclic redundancy check code, power supply status, and sensor diagnostic status.

[0048] Furthermore, the wireless radio frequency transceiver module 103 reads the transmission queue using interrupt or direct memory access methods, and completes wireless transmission according to at least one of the following strategies: node time slot, interleaved backoff, or event priority.

[0049] Furthermore, the tire pressure integrated sensing module 1 is installed inside the tire, on the rim, at the tire valve, or in a location connected to the tire's sealed cavity via an air guide structure. The air pressure acquisition module 101 communicates with the sealed cavity formed by the rim and the inflated tire through an air inlet or air guide channel, and the air inlet or air guide channel is equipped with a breathable and waterproof structure. The printed circuit board supporting the tire pressure integrated sensing module 1 has a digital ground area and a radio frequency ground area, and a copper-free clearance area is provided in the antenna radiation area of ​​the wireless radio frequency transceiver module 103.

[0050] Furthermore, the receiving and processing module 2 includes a dynamic thermal hysteresis compensation unit, an unloading state discrimination unit, a robust reference tracking unit, a noise scale estimation unit, a contact state calculation unit, and a morphological feature extraction unit. The robust reference tracking unit updates the dynamic reference only when the unloading state discrimination unit outputs a valid update gating signal, in order to prevent the pressure pulsation of the support phase from being written into the dynamic reference.

[0051] In practical applications, a tire pressure integrated sensor module is installed for each of the robot's front left wheel, front right wheel, rear left wheel, and rear right wheel. Each tire pressure integrated sensor module has a unique node identifier to distinguish different wheel positions and establishes a wireless communication connection with the receiving and processing module located on the robot body. The receiving and processing module is used to receive, verify, unpack, and time-align the data frames sent by each tire pressure integrated sensor module, and connects to the robot motion control host computer via a serial bus, Ethernet, controller area network, or other communication interface. Regarding the single-node hardware structure, such as... Figure 1 As shown, the tire pressure integrated sensing module adopts an integrated circuit structure, including a main control MCU module, a tire pressure acquisition module, a wireless radio frequency transceiver module, and a power management module.

[0052] In one specific implementation, the main control MCU module can be an STM32 series microcontroller, the barometric pressure acquisition module can be an MS5803-5BA digital pressure sensor, and the wireless RF transceiver module can be an E01 series wireless communication module operating in the 2.4 GHz band. The above device models are only used to illustrate one feasible hardware implementation; the main control MCU module can also use other microcontrollers with timing sampling, data processing, storage, and communication interface functions. The barometric pressure acquisition module can also use other pressure sensors that meet the corresponding pressure range, measurement accuracy, response speed, and temperature acquisition requirements. The wireless RF transceiver module can also use wireless communication devices operating in other licensed frequency bands that can meet the data refresh rate and communication latency requirements.

[0053] The air intake of the pressure acquisition module faces the sealed cavity formed by the flexible pneumatic tire and the rigid hub. The air intake can be covered with a breathable and waterproof membrane, and other circuits can be protected by potting, coating, or housing sealing. The power management module can be battery-powered or combined with energy acquisition or inductive power. Its outputs are connected to the pressure acquisition module, the main control MCU module, and the wireless RF transceiver module, respectively.

[0054] The raw digital air pressure (Dp) and raw digital temperature (Dt) acquired by the air pressure acquisition module are sent to the main control MCU module via SPI or I2C bus. After processing, the main control MCU module forms a data frame, which is then sent to the receiving and processing module by the wireless RF transceiver module. After completing secondary compensation and state calculation, the receiving and processing module sends at least one of the following to the robot motion control host computer: ground contact state, contact probability, contact confidence, secondary compensated air pressure, and terrain feature vector.

[0055] Furthermore, such as Figure 3 As shown, the main control MCU module runs non-blocking multi-rate acquisition firmware. The pressure sampling frequency can be set according to the robot control cycle, while the temperature sampling frequency is lower than the pressure sampling frequency. For example, a temperature sample can be inserted every few pressure sampling cycles. The frequencies, cycles, and data formats in the following steps can all be adjusted according to the sensor bandwidth, wireless link capability, and robot control cycle.

[0056] Step S1: After the system is powered on, the main control MCU module initializes the hardware timer, data bus, direct memory access unit, transmit queue and wireless radio frequency transceiver module, and reads the factory calibration coefficient from the calibration parameter storage unit of the air pressure acquisition module.

[0057] In step S2, a hardware timer triggers the pressure sampling task, and the air pressure acquisition module outputs the raw air pressure digital value Dp. The acquisition scheduling unit triggers the temperature sampling task in an interleaved manner according to a preset frequency division relationship to obtain the raw temperature digital value Dt. The sampling completion event is notified to the main control MCU module via interrupt or direct memory access, without using blocking polling that occupies the entire conversion cycle.

[0058] In step S3, the main control MCU module first converts Dp and Dt into physical pressure and temperature using the factory calibration coefficients, and then corrects the temperature-related zero-point error, sensitivity error, and nonlinearity error. Optionally, the node maintains the recursive thermal state, and estimates the thermal hysteresis component during rapid heating or cooling based on the current temperature, the rate of temperature change, and the thermal state at the previous moment, to obtain the first-level compensation pressure P1 (first-level sensor compensation).

[0059] In steps S4 and S5, the main control MCU module encapsulates at least one of the following into a data frame: node identifier, sampling time or relative timestamp, serial number, P1, temperature, power supply status, sensor diagnostic status, and cyclic redundancy check code, and writes it to the transmission queue. The wireless RF transceiver module performs asynchronous burst transmission via direct memory access or a transmission completion interrupt. During transmission, the next pressure sampling cycle can still continue execution according to the hardware timer.

[0060] Furthermore, after receiving a data frame, the receiving and processing module first determines the frame integrity based on the Cyclic Redundancy Check (CRC) code, assigns the data to the corresponding wheel and foot sequence based on the node identifier, marks lost packets based on the sequence number, and resamples multiple wheel and foot data onto a unified time axis based on the time information. For short-term missing data, it can be marked as missing, or it can be filled in using preservation, interpolation, or model prediction, but the filled data is not used to update the dynamic baseline.

[0061] Let the first-level compensation pressure be P1(k), the dynamic thermal hysteresis estimate be Pth(k), and the dynamic baseline be b(k). Then, the second-level compensation pressure can be expressed as Pc(k) = P1(k) - Pth(k) - b(k). Here, Pth(k) is recursively obtained from the current temperature, the rate of temperature change, and the thermal state at the previous moment. It is used to characterize the asynchronous thermal response (second-level dynamic drift compensation) between the sensor chip, tire gas, and mounting structure when the temperature changes rapidly.

[0062] The dynamic baseline b(k) is not updated at all times. The receiving and processing module constructs an unloading state gating signal G(k): when the absolute value of the rate of change of air pressure is lower than the rate of change threshold, the sliding window variance is lower than the variance threshold, the contact probability is lower than the probability threshold, and the above conditions are continuously met for a preset duration, G(k) is set to 1. Otherwise, G(k) is set to 0. Only when G(k) = 1 is the dynamic baseline updated with the statistics of the current unloading window.

[0063] In one embodiment, the median of P1(k)-Pth(k) within the unloading window is used as a candidate benchmark, the noise scale sigma is estimated using the median absolute deviation (MAD), and a finite update rate is employed to allow b(k) to slowly track ambient air pressure, slow leaks, and long-term zero drift. The median and MAD are insensitive to ground impacts, wireless outliers, and a small number of erroneous samples. The finite update rate prevents short-term temperature fluctuations from causing abrupt changes in the benchmark.

[0064] When the wheel enters the support phase or a high rate of change pressure pulse occurs, G(k)=0, and the dynamic reference remains unchanged, thus preventing the actual ground contact pressure from being written into the reference. If the gating condition cannot be met for several consecutive unloading cycles, the previous valid reference can be maintained and the output confidence level reduced, or the reference update can be restarted by combining the swing phase prior provided by the robot motion control host computer.

[0065] Furthermore, the receiving and processing module constructs a feature vector from Pc(k). This feature vector may include the second-order compensated pressure increment, the first-order rate of change of pressure, the second-order rate of change of pressure, the sliding window energy, the duration, and the current noise scale. The contact state calculation unit outputs the contact probability pc(k) through logistic regression, a Bayesian model, a state machine, a hidden Markov model, or a calibrated nonlinear mapping. This embodiment does not limit the specific model of the contact probability calculator, but its input must at least simultaneously include pressure amplitude information and dynamic change information.

[0066] The ground contact trigger threshold Hon and the ground release threshold Hoff are adaptively determined based on at least two of the following: noise scale sigma, static tire pressure, temperature state, and movement speed, and Hon > Hoff. When pc(k) or the corresponding ground contact criterion exceeds Hon, the system enters the support phase. When it falls below Hoff for a preset duration, the system enters the swing phase. The dual-threshold hysteresis structure suppresses frequent flipping caused by noise near the thresholds, and the threshold changes with sigma, automatically increasing the trigger requirement (ground contact event segmentation) when noise increases.

[0067] Contact confidence can be determined by the contact probability, a distance of 0.5, current data integrity, baseline validity, rate of temperature change, and model residuals. The robot motion control host computer can adjust the observation noise, gait switching conditions, or control gain in the state estimation based on the contact confidence, instead of unconditionally treating the binary ground contact state as the absolute truth (transient morphological feature extraction).

[0068] Furthermore, after a ground contact event is segmented by a start time (ton) and an end time (toff), the event time is mapped to a normalized time axis of 0 to 1, and the Pc waveform is normalized by the event peak value, event energy, or steady-state pressure increment. Subsequently, features such as rise time, peak position, loading slope, unloading slope, ratio of loading area to unloading area, peak attenuation ratio, kurtosis, and preset frequency band energy are extracted. Since the features are based on the relative morphology of the normalized waveform, the direct impact of robot mass, instantaneous landing force, initial tire pressure, and movement speed on terrain recognition can be reduced (transient morphological feature extraction).

[0069] The receiving and processing module arranges the contact probabilities of multiple wheel feet along a unified time axis to form a contact pattern vector. It then concatenates, statistically aggregates, or weights the terrain feature vectors of each wheel foot according to contact confidence to form a comprehensive terrain description of the robot. The robot's motion control host computer can use the contact pattern vector for state estimation and gait phase correction, and the terrain description for terrain classification, gait parameter adjustment, or as observation input for reinforcement learning strategies.

[0070] During implementation, calibration tests can be used to obtain the first-level compensation coefficient, thermal state model parameters, gating threshold, adaptive threshold coefficient, and terrain classifier parameters. Calibration data should cover different temperatures, initial tire pressures, loads, movement speeds, and typical terrains. However, during online operation, the dynamic baseline, noise scale, and contact threshold are still adaptively updated based on current data to avoid the system being applicable only to a single calibration condition.

[0071] An embodiment of the present invention provides a ground contact sensing method applied to the aforementioned wheeled robot tire pressure integrated sensing system, comprising: S1, a pressure acquisition module acquiring raw digital air pressure and raw digital temperature; S2, a main control MCU module performing first-level temperature, nonlinearity, and dynamic thermal hysteresis compensation on the raw digital air pressure to obtain first-level compensated air pressure data; S3, the main control MCU module encapsulating the first-level compensated air pressure data with node identifiers, time information, and sequence numbers into a data frame, and transmitting it to a receiving and processing module via a wireless radio frequency transceiver module; S4, the receiving and processing module performing integrity verification, packet loss marking, and time alignment on the data frame; S5, the receiving and processing module updating the robust dynamic reference and noise scale according to the unloading state gating, and subtracting the dynamic thermal hysteresis component and dynamic reference from the first-level compensated air pressure data to obtain second-level compensated air pressure data; S6, calculating the contact probability based on the second-level compensated air pressure data, air pressure change rate, and noise scale, and using an adaptive dual threshold to determine the ground contact state for ground contact sensing. S7 performs amplitude and time normalization on the secondary compensation air pressure waveform of a single ground contact event, extracts terrain feature vectors, and outputs the ground contact perception results containing ground contact state, contact confidence and terrain feature vectors to the robot motion control host computer.

[0072] Furthermore, the dynamic thermal hysteresis compensation in step S2 employs a recursive thermal state model. The inputs to the recursive thermal state model include at least the current temperature, the first-order rate of change of temperature, and the thermal state at the previous moment. The output consists of the slowly varying pressure component caused by temperature changes. The slowly varying pressure component is separated from the rapidly varying pressure component caused by ground contact at different time scales.

[0073] Furthermore, the unloading state gating conditions in step S5 include at least: the absolute value of the rate of change of air pressure is lower than the rate of change threshold, the variance of the sliding window is lower than the variance threshold, and the contact probability is lower than the probability threshold. The dynamic benchmark is updated using the median of the sliding window and the noise scale is updated using the median absolute deviation, provided that the unloading state gating conditions are continuously met for a preset duration.

[0074] Furthermore, the ground contact trigger threshold and ground release threshold in step S6 are adaptively determined by at least two of the following: noise scale, tire static pressure, temperature state, and movement speed. The contact probability is calculated based on at least three of the following: secondary compensated air pressure data, first-order air pressure change rate, second-order air pressure change rate, window energy, and duration, and the ground contact state is output through hysteresis logic where the trigger threshold is higher than the release threshold.

[0075] Further, step S7 includes: normalizing the waveform time axis with the start and end times of the ground contact event, normalizing the waveform amplitude with peak value, energy or steady-state increment, and extracting at least four items from the normalized waveform, including rise time, peak position, loading slope, unloading slope, area ratio, attenuation ratio, kurtosis and preset frequency band energy, to form a terrain feature vector that weakens the influence of single-step stamping force and movement speed.

[0076] Furthermore, the receiving and processing module 2 performs unified time axis resampling on the data of multiple wheel feet based on the time information in the data frames of each tire pressure integrated sensing module 1, constructs a contact pattern vector composed of the contact probabilities of each wheel foot, and sends the contact pattern vector and the terrain feature vector together to the robot motion control host computer 3 for state estimation, gait planning or control strategy switching.

[0077] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the preferred embodiments, while those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0078] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A tire pressure integrated sensing system for a wheeled robot, characterized in that, It includes a tire pressure integrated sensing module connected to the air passage of the robot's flexible pneumatic tire, a receiving and processing module set on the robot body, and a robot motion control host computer that is communicatively connected to the receiving and processing module. The tire pressure integrated sensing module includes a pressure acquisition module, a main control MCU module, a wireless radio frequency transceiver module, and a power management module. The air pressure acquisition module is used to acquire the original digital air pressure and the original digital temperature in the sealed cavity of the tire, and send the original digital air pressure and the original digital temperature to the main control MCU module through the data bus. The main control MCU module is used to perform acquisition timing scheduling, first-level sensor compensation and data frame encapsulation, generate data frames, and send the data frames to the wireless radio frequency transceiver module. The wireless radio frequency transceiver module is used to wirelessly transmit the data frame to the receiving and processing module. The receiving and processing module is used to perform integrity verification and time alignment on the data frame, perform secondary dynamic drift compensation, ground contact event segmentation, contact state calculation and transient morphological feature extraction, and output ground contact state, contact confidence and terrain feature vector to the robot motion control host computer. The power management module is electrically connected to and supplies power to the air pressure acquisition module, the main control MCU module, and the wireless radio frequency transceiver module.

2. The wheel-legged robot tire pressure integrated sensing system according to claim 1, characterized in that, The air pressure acquisition module includes a pressure sensing unit, a temperature detection unit, an analog-to-digital conversion unit, and a calibration parameter storage unit. The air pressure acquisition module communicates with the main control MCU module via an SPI bus or an I2C bus, and outputs the raw air pressure digital value at a pressure sampling frequency higher than the temperature sampling frequency.

3. The wheel-legged robot tire pressure integrated sensing system according to claim 2, characterized in that, The main control MCU module includes a hardware timer triggering unit, a direct memory access unit, an acquisition scheduling unit, and a transmission queue. The acquisition scheduling unit adopts a multi-rate scheduling method of interleaved high-frequency pressure sampling and low-frequency temperature sampling, and decouples the air pressure acquisition task, compensation calculation task, and wireless transmission task so that any wireless transmission cycle does not block the next pressure sampling cycle. The primary sensor compensation includes: performing physical quantity calculation on the original digital air pressure based on the calibration coefficients in the calibration parameter storage unit; correcting zero point, sensitivity, and nonlinearity errors based on temperature data; and estimating the dynamic thermal hysteresis component between the pressure sensing unit and the tire cavity based on the current temperature, temperature change rate, and historical temperature status to obtain primary compensated air pressure data.

4. The wheel-legged robot tire pressure integrated sensing system according to claim 1, characterized in that, The data frame includes at least one of the following: node identifier, time information, sequence number, compensated air pressure data, temperature data, sampling time, sampling interval, cyclic redundancy check code, power supply status, and sensor diagnostic status; the wireless radio frequency transceiver module reads the transmission queue using interrupt or direct memory access methods, and completes wireless transmission according to at least one of the following strategies: node time slot, staggered backoff, or event priority.

5. The wheel-legged robot tire pressure integrated sensing system according to claim 1, characterized in that, The tire pressure integrated sensing module is installed inside the tire, on the rim, at the tire valve, or in a location connected to the tire's sealed cavity via an air guide structure. The air pressure acquisition module is connected to the sealed cavity formed by the rim and the inflatable tire via an air inlet or air guide channel, and the air inlet or air guide channel is equipped with a breathable and waterproof structure. The printed circuit board supporting the tire pressure integrated sensing module has a digital ground area and a radio frequency ground area, and a copper-free clearance area is set in the antenna radiation area of ​​the wireless radio frequency transceiver module.

6. The wheel-legged robot tire pressure integrated sensing system according to claim 1, characterized in that, The receiving and processing module includes a dynamic thermal hysteresis compensation unit, an unloading state discrimination unit, a robust benchmark tracking unit, a noise scale estimation unit, a contact state calculation unit, and a morphological feature extraction unit. The robust reference tracking unit updates the dynamic reference only when the unloading state discrimination unit outputs a valid update gating signal, so as to avoid the support phase pressure pulsation being written into the dynamic reference.

7. A ground-sensing method, characterized in that, The wheel-footed robot tire pressure integrated sensing system according to any one of claims 1 to 6 includes: S1, a pressure acquisition module acquires raw digital air pressure and raw digital temperature; S2, a main control MCU module performs first-level temperature, nonlinearity, and dynamic thermal hysteresis compensation on the raw digital air pressure to obtain first-level compensated air pressure data; S3, the main control MCU module encapsulates the first-level compensated air pressure data with node identifier, time information, and sequence number into a data frame, and transmits it to the receiving and processing module through a wireless radio frequency transceiver module; S4, the receiving and processing module performs integrity verification, packet loss marking, and time alignment on the data frame. S5, the receiving and processing module updates the robust dynamic reference and noise scale according to the unloading state gating, and subtracts the dynamic thermal hysteresis component and dynamic reference from the primary compensated air pressure data to obtain the secondary compensated air pressure data; S6, the contact probability is calculated based on the secondary compensated air pressure data, air pressure change rate and noise scale, and the ground contact state is determined by adaptive dual thresholds for ground contact sensing; S7, the amplitude and time of the secondary compensated air pressure waveform of a single ground contact event are normalized, the terrain feature vector is extracted, and the ground contact sensing result containing the ground contact state, contact confidence and terrain feature vector is output to the robot motion control host computer.

8. The ground contact sensing method according to claim 7, characterized in that, The dynamic thermal hysteresis compensation in step S2 adopts a recursive thermal state model. The input of the recursive thermal state model includes at least the current temperature, the first-order rate of change of temperature and the thermal state at the previous moment, and the output is the slowly changing pressure component caused by the temperature change. The slowly varying air pressure component and the rapid air pressure component caused by ground contact are separated at different time scales; The unloading state gating conditions in step S5 include at least the following: the absolute value of the rate of change of air pressure is lower than the rate of change threshold, the variance of the sliding window is lower than the variance threshold, and the contact probability is lower than the probability threshold. The dynamic benchmark is updated using the median of the sliding window and the noise scale is updated using the median absolute deviation, only if the unloading state gating condition continues to meet the preset duration.

9. The ground contact sensing method according to claim 7, characterized in that, The ground contact trigger threshold and ground release threshold in step S6 are adaptively determined by at least two of the noise scale, tire static pressure, temperature state, and speed of movement. The contact probability is calculated based on at least three of the following: secondary compensated air pressure data, first-order rate of change of air pressure, second-order rate of change of air pressure, window energy, and duration, and the ground contact state is output by hysteresis logic when the trigger threshold is higher than the release threshold. Step S7 includes: normalizing the waveform time axis with the start and end times of the ground contact event, normalizing the waveform amplitude with peak value, energy or steady-state increment, and extracting at least four items from the normalized waveform, including rise time, peak position, loading slope, unloading slope, area ratio, attenuation ratio, kurtosis and preset frequency band energy, to form a terrain feature vector that weakens the influence of single-step stamping force and movement speed.

10. The ground contact sensing method according to claim 7, characterized in that, The receiving and processing module performs unified time axis resampling on the data of multiple wheel feet based on the time information in the data frames of each tire pressure integrated sensing module, constructs a contact pattern vector composed of the contact probabilities of each wheel foot, and sends the contact pattern vector and the terrain feature vector together to the robot motion control host computer for state estimation, gait planning or control strategy switching.

Images