# US 2021/0331677 A1 · Vehicle control apparatus

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0331677 A1
- Document ID: US20210331677A1
- Status: Published application
- Kind: A1
- Application number: 16/494,726
- Filed: 2019-05-29
- Published: 2021-10-28
- Earliest claimed: 2019-05-29
- Assignees: LG Electronics Inc.
- Applicants: LG Electronics Inc.
- Inventors: Taehyun KIM
- Web page: https://patentgrep.com/patent/US20210331677A1

## CPC

- B60W 30/146: {Speed limiting}
- B60W 30/18009: {related to particular drive situations}
- B60W 10/04: including control of propulsion units
- B60W 10/18: including control of braking systems
- B60W 40/10: related to vehicle motion
- B60W 40/13: Load or weight
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/092: Reinforcement learning
- B60W 2050/0008: {Feedback, closed loop systems or details of feedback error signal}
- B60W 2510/22: Suspension systems
- B60W 2520/10: Longitudinal speed
- B60W 2530/10: Weight
- B60W 2540/10: Accelerator pedal position
- B60W 2540/12: Brake pedal position
- B60W 2552/00: Input parameters relating to infrastructure
- B60W 2552/35: Road bumpiness, e.g. potholes
- B60W 2554/20: Static objects
- B60W 2710/182: Brake pressure, e.g. of fluid or between pad and disc
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G06N 3/047: Probabilistic or stochastic networks
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- 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

A vehicle control apparatus is disclosed. The vehicle control apparatus includes a distance sensor configured to sense a distance from a speed bump, a speed sensor configured to sense a speed of a vehicle, a vehicle driver configured to control an operation of a brake or an operation of an accelerator, and a processor configured to provide an input parameter comprising a speed and a weight of the vehicle to a prediction model and control the vehicle driver so that the brake is operated at a braking intensity outputted from the prediction model when the vehicle enters the speed bump on the basis of the distance from the speed bump.

## Claims

### Claim 1

A vehicle control apparatus comprising:
a distance sensor configured to sense a distance from a speed bump;
a speed sensor configured to sense a speed of a vehicle;
a vehicle driver configured to control an operation of a brake or an operation of an accelerator; and
a processor configured to provide an input parameter comprising a speed and a weight of the vehicle to a prediction model and control the vehicle driver so that the brake is operated at a braking intensity outputted from the prediction model when the vehicle enters the speed bump on the basis of the distance from the speed bump.

### Claim 2 (depends on 1)

The vehicle control apparatus according to claim 1, wherein the prediction model is generated by, based on reinforcement learning, allowing a neural network to receive a training input parameter so as to output the braking intensity and to be trained based on feedback on the braking intensity outputted by the neural network,
wherein the training input parameter comprises at least one of the speed of the vehicle, the weight of the vehicle, or a height of the speed pump, and
the feedback is vertical movement after the vehicle enters the speed bump.

### Claim 3 (depends on 2)

The vehicle control apparatus according to claim 2, wherein the prediction model is generated by allowing the neural network to be trained using at least one of reward on the reduction of the vertical movement of the vehicle or penalty on the increase of the vertical movement of the vehicle.

### Claim 4 (depends on 3)

The vehicle control apparatus according to claim 3, wherein the prediction model acquires policy for minimizing the vertical movement of the vehicle after entering the speed bump.

### Claim 5 (depends on 1)

The vehicle control apparatus according to claim 1, wherein the processor operates the brake at the braking intensity outputted from the prediction model when a front wheel of the vehicle enters the speed bump.

### Claim 6 (depends on 1)

The vehicle control apparatus according to claim 1, wherein the input parameter comprises the weight comprising an unladen vehicle weight and an additional loading weight.

### Claim 7 (depends on 1)

The vehicle control apparatus according to claim 1, wherein the input parameter further comprises a height of the speed bump.

### Claim 8 (depends on 1)

The vehicle control apparatus according to claim 1, wherein the processor is configured to control the vehicle driver configured to provide a second input parameter comprising a second speed and the weight of the vehicle to a second prediction model and operate the accelerator at an acceleration intensity outputted from the second prediction model when a rear wheel of the vehicle enters the speed bump.

### Claim 9 (depends on 8)

The vehicle control apparatus according to claim 8, wherein the second prediction model is generated by, based on reinforcement learning, allowing a second neural network to receive a training input parameter so as to output the acceleration intensity and to be trained based on second feedback on the acceleration intensity outputted by the second neural network,
wherein the training input parameter comprises at least one of the speed of the vehicle, the weight of the vehicle, or a height of the speed pump, and
the second feedback is vertical movement after the rear wheel of the vehicle enters the speed bump.

### Claim 10 (depends on 9)

The vehicle control apparatus according to claim 9, wherein the second prediction model is generated by allowing the second neural network to be trained using at least one of reward on the reduction of the vertical movement of the vehicle or penalty on the increase of the vertical movement of the vehicle.

### Claim 11 (depends on 10)

The vehicle control apparatus according to claim 10, wherein the second prediction model acquires second policy for minimizing the vertical movement of the vehicle after the rear wheel enters the speed bump.

### Claim 12

A method for generating a prediction model, the method comprising:
providing a training input parameter, which comprises a speed and a weight of a vehicle, to a neural network;
acquiring feedback on a braking intensity when the neural network outputs the braking intensity; and
training the neural network in a reinforcement learning manner on the basis of the feedback,
wherein the feedback is vertical movement after the vehicle enters a speed bump.

### Claim 13 (depends on 12)

The method according to claim 12, wherein the weight of the vehicle comprises an unladen vehicle weight and an additional loading weight.

### Claim 14 (depends on 12)

The method according to claim 12, wherein the training input parameter further comprises a height of the speed bump.

### Claim 15 (depends on 12)

The method according to claim 12, further comprising:
providing a second training input parameter, which comprises the speed and the weight of the vehicle, to a second neural network;
acquiring second feedback on an acceleration intensity when the second neural network outputs the acceleration intensity; and
training the second neural network in a reinforcement learning manner on the basis of the second feedback,
wherein the second feedback is vertical movement after a rear wheel of the vehicle enters a speed bump.
