Measuring apparatus and method

1. A measuring apparatus comprising:

an ultrasonic transmitter that controls transmission of ultrasonic waves from a plurality of ultrasonic transmitting devices to form an ultrasonic beam having directivity and transmits, as a transmission signal, the ultrasonic beam in a plurality of directions;

an ultrasonic receiver that receives, as received signals, reflected waves of the transmission signal from the plurality of directions, one received signal including a plurality of reflected waves when the transmission signal is transmitted to one direction of the plurality of directions, the received signals each representing a waveform including a plurality of local maximum values; and

an estimator that estimates range information from the received signals, based on a multi-layer neural network which is trained using the received signals as input data, and preliminarily obtained received signals and a preliminarily obtained distance to an object as answer data, the multi-layer neural network including a plurality of intermediate layers, each of which includes a plurality of nodes, wherein the nodes are coupled to one another only between different intermediate layers.

2. The apparatus according to claim 1 , wherein the ultrasonic transmitter transmits the transmission signal while performing a scanning operation in the plurality of directions so that adjacent portions of the ultrasonic beam overlap.

3. The apparatus according to claim 2 , wherein each of the plurality of directions is an azimuth direction in a two-dimensional plane.

4. The apparatus according to claim 2 , wherein the plurality of directions is an azimuth direction and a direction of an elevation angle in a three-dimensional space.

5. The apparatus according to claim 1 , further comprising:

a plurality of bandpass filters that generate a plurality of filtered received-signals from the received signals by passing a signal having a specific frequency band,

wherein the estimator estimates the range information from the plurality of filtered received-signals, based on a plurality of preliminarily obtained filtered received-signals and a distance to the object before and after measuring using an ultrasonic beam.

6. The apparatus according to claim 1 , further comprising:

a one-dimensional range sensor that measures a distance from the apparatus to an object in a given direction to obtain one-dimensional range information,

wherein the estimator estimates the range information from the one-dimensional information and the received signals, based on preliminarily obtained reflected waves from the plurality of directions, preliminarily obtained one-dimensional range information and a preliminarily obtained distance to the object.

7. The apparatus according to claim 1 , further comprising:

an analyzer that performs machine learning using the received signals and multi-dimensional range information related to a distance from the apparatus to an object, and generates a learning result including an estimation formula, the multi-dimensional range information being obtained by a multi-dimensional range sensor.

8. The apparatus according to claim 1 , wherein the estimator estimates the range information, based on a learning result related to the preliminarily obtained received signals and the preliminarily obtained distance to the object.

9. A measuring method comprising:

controlling transmission of ultrasonic waves from a plurality of ultrasonic transmitting devices to form an ultrasonic beam having directivity;

transmitting, as a transmission signal, the ultrasonic beam in a plurality of directions;

receiving, as received signals, reflected waves of the transmission signal from the plurality of directions, one received signal including a plurality of reflected waves when the transmission signal is transmitted to one direction of the plurality of directions, the received signals each representing a waveform including a plurality of local maximum values; and

performing machine learning on multi-layer neural network which is trained using the received signals as input data and multi-dimensional range information related to a distance from a measuring apparatus to an object as answer data, to generate a learning result including an estimation formula, the multi-dimensional range information being obtained by a multi-dimensional range sensor, the multi-layer neural network including a plurality of intermediate layers, each of which includes a plurality of nodes, wherein the nodes are coupled to one another only between different intermediate layers.

10. A measuring method comprising:

controlling transmission of ultrasonic waves from a plurality of ultrasonic transmitting devices to form an ultrasonic beam having directivity;

transmitting, as a transmission signal, the ultrasonic beam in a plurality of directions;

receiving, as received signals, reflected waves of the transmission signal from the plurality of directions, one received signal including a plurality of reflected waves when the transmission signal is transmitted to one direction of the plurality of directions, the received signals each representing a waveform including a plurality of local maximum values; and

estimating range information from the received signals, based on a multi-layer neural network which is trained using the received signals as input data, and preliminarily obtained received signals and a preliminarily obtained distance to an object as answer data, the multi-laver neural network including a plurality of intermediate layers, each of which includes a plurality of nodes, wherein the nodes are coupled to one another only between different intermediate layers.