# US 2026/0097776 A1 · Apparatus and method of diagnosing a cause of increased exhaust gas using an autoencoder, and a vehicle including the same > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2026/0097776 A1 - Document ID: US20260097776A1 - Status: Published application - Kind: A1 - Application number: 19/175,547 - Filed: 2025-04-10 - Published: 2026-04-09 - Earliest claimed: 2024-10-04 - Assignees: Hyundai Motor Company; Kia Corporation - Applicants: Hyundai Motor Company; Kia Corporation - Inventors: Jin Gwon Park - Web page: https://patentgrep.com/patent/US20260097776A1 ## CPC - B60W 50/0205: {Diagnosing or detecting failures; Failure detection models} - G06N 3/0455: Auto-encoder networks; Encoder-decoder networks - B60W 2050/021: {Means for detecting failure or malfunction} - B60W: CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract An apparatus for diagnosing a cause of an increase in exhaust gas using an autoencoder includes: at least one processor; and a storage medium for storing a computer-readable instruction, wherein when the computer-readable instruction is executed by the at least one processor, the computer-readable instruction is configured to cause the processor to: receive input data, wherein the input data includes at least two or more factors including a sensed value; determine a reconstruction error for each of the at least two or more factors, wherein the reconstruction error is a reconstruction error between the input data and restored data obtained by inputting the input data to a pre-learned autoencoder; and diagnose the cause of the increase in exhaust gas, based on the reconstruction errors. ## Claims ### Claim 1 An apparatus for diagnosing a cause of an increase in exhaust gas using an autoencoder, the apparatus comprising: at least one processor; and a storage medium configured to store a computer-readable instruction, wherein when the computer-readable instruction is executed by the at least one processor, the computer-readable instruction is configured to cause the at least one processor to: receive input data, wherein the input data includes at least two or more factors including a sensed value; determine a reconstruction error for each of the at least two or more factors, wherein the reconstruction error is a reconstruction error between the input data and restored data obtained by inputting the input data to a pre-learned autoencoder; and diagnose the cause of the increase in exhaust gas, based on each of the reconstruction errors. ### Claim 2 (depends on 1) The apparatus of claim 1, wherein the at least one processor is configured to accumulate and average the reconstruction errors to get an accumulated and averaged reconstruction error. ### Claim 3 (depends on 2) The apparatus of claim 2, wherein the at least one processor is further configured to: determine an error rate for each of the at least two or more factors, wherein the error rate is a value obtained by dividing the accumulated and averaged reconstruction error by a preset reference value; and diagnose the cause of the increase in exhaust gas, based on each of the error rates. ### Claim 4 (depends on 3) The apparatus of claim 3, wherein the preset reference value is an average value of the reconstruction errors obtained in a normal driving state in which an exhaust gas is lower than or equal to a preset value for each of the at least two or more factors. ### Claim 5 (depends on 2) The apparatus of claim 2, wherein the at least one processor is further configured to: determine whether a measured nitrogen oxide (NOx) concentration value exceeds a preset value; and diagnose the cause of the increase in exhaust gas when the measured NOx concentration value exceeds the preset value, or when the measured NOx concentration value exceeds the preset value and a warning light is turned on. ### Claim 6 (depends on 1) The apparatus of claim 1, wherein the at least one processor is configured to determine whether a stable driving condition is satisfied, and wherein the stable driving condition includes one or more of a vehicle speed, an engine revolutions per minute(RPM)value, a state in which a reverse gear is not engaged, an exhaust gas recirculation (EGR) ratio, a state in which a lambda control factor is not 1, a state in which learning of an on-board diagnostics (OBD) item is completed, an atmospheric pressure, or a state in which a driving mode of a vehicle is not a sports mode. ### Claim 7 (depends on 3) The apparatus of claim 3, wherein the at least two or more factors comprises one or more of a fuel pressure, a rear lambda voltage, a lambda control factor, a variable valve timing (VVT) incam error, a VVT excam error, a long term fuel trim (LTFT), an engine roughness sum, an engine gas recirculation (EGR) flow error, or a variable geometry turbocharger (VGT) value error. ### Claim 8 (depends on 7) The apparatus of claim 7, wherein the at least one processor is further configured to diagnose, as the cause of the increase in exhaust gas: a fuel system when an error rate of the fuel pressure is greater than a preset error rate, a rear oxygen sensor when an error rate of the rear lambda voltage is greater than the preset error rate, a front oxygen sensor when an error rate of the lambda control factor is greater than the preset error rate, an intake VVT when an error rate of the VVT incam error is greater than the preset error rate, an exhaust VVT when an error rate of the VVT excam error is greater than the preset error rate, an air-fuel ratio when an error rate of the long term fuel trim is greater than the preset error rate, a misfire when an error rate of the engine roughness sum is greater than the preset error rate, an EGR when an error rate of the EGR flow error is greater than the preset error rate, and a VGT when an error rate of the VGT valve error is greater than the preset error rate. ### Claim 9 (depends on 8) The apparatus of claim 8, wherein, when the air-fuel ratio is diagnosed as the cause of the increase in exhaust gas, the at least one processor is configured to diagnose the air-fuel ratio as a lean state when the air-fuel ratio is greater than 1, or diagnose the air-fuel ratio as a rich state when the air-fuel ratio is less than 1. ### Claim 10 (depends on 1) The apparatus of claim 1, wherein the input data further comprises engine status data, and wherein the engine status data includes one or more of a map sensor signal, an injection time, an engine revolution per minute (RPM) value, an engine load, a vehicle speed, a variable valve timing (VVT) incam angle, a VVT excam angle, a lambda controller output, a front lambda voltage, a lambda set value or a controller value. ### Claim 11 (depends on 3) The apparatus of claim 3, wherein the at least one processor is further configured to output the cause of the increase in exhaust gas. ### Claim 12 (depends on 3) The apparatus of claim 3, wherein the at least one processor is further configured to: output a cause of an increase related to a factor in magnitude order of an error rate, for each of the at least two or more factors, or output a cause of an increase related to a factor of a preset number of which a magnitude of an error rate satisfies a threshold, among the at least two or more factors, in magnitude order of an error rate. ### Claim 13 A method for diagnosing a cause of an increase in exhaust gas using an autoencoder, the method comprising: receiving input data, wherein the input data includes at least two or more factors including a sensed value; determining a reconstruction error for each of the at least two or more factors, wherein the reconstruction error is a reconstruction error between the input data and restored data obtained by inputting the input data to a pre-learned autoencoder; and diagnosing the cause of the increase in exhaust gas, based on each of the reconstruction errors. ### Claim 14 (depends on 13) The method of claim 13, further comprising accumulating and averaging the reconstruction errors to get an accumulated and averaged reconstruction error. ### Claim 15 (depends on 14) The method of claim 14, wherein diagnosing the cause of the increase in exhaust gas comprises: determining an error rate for each of the at least two or more factors, wherein the error rate is a value obtained by dividing the accumulated and averaged reconstruction error by a preset reference value; and diagnosing the cause of the increase in exhaust gas, based on each of the error rates. ### Claim 16 (depends on 15) The method of claim 15, wherein the preset reference value is an average value of the reconstruction errors obtained in a normal driving state in which an exhaust gas is lower than or equal to a preset value for each of the at least two or more factors. ### Claim 17 (depends on 14) The method of claim 14, further comprising determining whether a measured nitrogen oxide (NOx) concentration value exceeds a preset value, and wherein diagnosing the cause of the increase in exhaust gas is performed when the measured NOx concentration value exceeds the preset value, or when the measured NOx concentration value exceeds the preset value and a warning light is turned on. ### Claim 18 (depends on 15) The method of claim 15, wherein the at least two or more factors comprises one or more of a fuel pressure, a rear lambda voltage, a lambda control factor, a variable valve timing (VVT) incam error, a VVT excam error, a long term fuel trim (LTFT), an engine roughness sum, an engine gas recirculation (EGR) flow error, or a variable geometry turbocharger (VGT) value error. ### Claim 19 (depends on 18) The method of claim 18, wherein diagnosing the cause of the increase in exhaust gas further comprises diagnosing, as the cause of the increase in exhaust gas: a fuel system when an error rate of the fuel pressure is greater than a preset error rate, a rear oxygen sensor when an error rate of the rear lambda voltage is greater than the preset error rate, a front oxygen sensor when an error rate of the lambda control factor is greater than the preset error rate, an intake VVT when an error rate of the VVT incam error is greater than the preset error rate, an exhaust VVT when an error rate of the VVT excam error is greater than the preset error rate, an air-fuel ratio when an error rate of the long term fuel trim is greater than the preset error rate, a misfire when an error rate of the engine roughness sum is greater than the preset error rate, an EGR when an error rate of the EGR flow error is greater than the preset error rate, and a VGT when an error rate of the VGT valve error is greater than the preset error rate. ### Claim 20 A vehicle including an apparatus for diagnosing a cause of an increase in exhaust gas using an autoencoder, wherein the apparatus comprises: at least one processor; and a storage medium configured to store a computer-readable instruction, wherein when the computer-readable instruction is executed by the at least one processor, the computer readable instruction is configured to cause the at least one processor to: receive input data, wherein the input data includes at least two or more factors including a sensed value; determine a reconstruction error for each of the at least two or more factors, wherein the reconstruction error is a reconstruction error between the input data and restored data obtained by inputting the input data to a pre-learned autoencoder; and diagnose the cause of the increase in exhaust gas, based on each of the reconstruction errors.