JP3866684B2 - Radar signal processing apparatus and processing method - Google Patents

Description

【００３９】
図９はこの式で用いられる変数ｉとｊの説明図である。同図においてｉは図７に対応する横軸の座標の値に、またｊは縦軸に対応する座標の値に相当し、図７に対応してｉの値は０から５、ｊの値は０から３の範囲となる。
【００４０】
図８のステップＳ１２で、ある格子点から隣接する格子点までの経路のうち、最小のコストの値の検出とその蓄積が行なわれる。ステップＳ１３では最小コストとなる隣接格子点の位置情報が蓄積され、ステップＳ１４で終点位置に達したと判定されるまでステップＳ１２以降の処理が繰り返され、終点位置に達した時点でＤＰマッチング処理を終了する。
【００４１】
ステップＳ１２の最小コスト値検出処理においては、次式を用いて計算が行なわれる。
【００４２】
【数２】

【００４３】
ここでｋは０から始まり、ステップＳ１２〜Ｓ１４の繰り返しの回数に相当し、Ｆは図７で説明した各方向の重みである。ｄはピーク受信電力の対数値の差の絶対値であり、ｇは１つの格子点から隣接する３つの格子点までのコストのうちどれが最小になるかを算出する時の、その１つの格子点までの探索された経路に対応するコストの値である。３つの値のうち最小のものが検出され、その最小コストの値に対応する隣接格子点の位置が、座標を示すｉ（ｋ＋１），ｊ（ｋ＋１）に代入され、ステップＳ１３でその位置が蓄積される。
【００４４】
図１０は、最も左下の格子点（Ｕ０,Ｄ０）から、３つの隣接格子点のうち、どの方向に進むべきかを決定する最小コスト値検出処理の説明図である。この処理において上方向、上斜め方向、および横方向に進む場合のｇの値は次式によって計算される。
【００４５】
【数３】

【００４６】
この３つのうち、横方向に進むコストｇ（１，０）が最小となるため、図１０に示すように横方向に進む経路が最適経路の一部として決定される。
図８のステップＳ１２、Ｓ１３の処理を繰り返すことによって、図１０の状態から先の最適経路が次々と決定される。図１１はその結果を示す。この結果から、図５の上昇区間における３つのピークＵ０〜Ｕ２は、図６の下降区間における１つのピークＤ０と対応し、上昇区間における３つのピークＵ３〜Ｕ５は、下降区間における３つのピークＤ１〜Ｄ３にそれぞれ１対１に対応するというマッチング結果が得られる。
【００４７】
図１２はＤＰマッチングの第２の方法の処理フローチャートである。この第２の方法においては、図７における各格子点、すなわち２４個の格子点のそれぞれに対して、スタート点からその格子点に至るまでの経路の中でコストが最小となる経路のコストを各格子位置の最小コスト値として検出した後に、エンド点側から最小コストとなる格子点の位置を追跡し、スタート点まで至り、最適経路を探索するものである。
【００４８】
図１２において処理が開始されると、まずステップＳ１６で図８のステップＳ１１と同様にｇ（０，０）、すなわちスタート点から（Ｕ０，Ｄ０）の格子点まで達するためのコストの値が計算された後に、ステップＳ１７で前述のようにそれぞれの格子点に対応する最小コストの値の検出と蓄積が行なわれ、ステップＳ１８で計算が行なわれた格子点の位置から見て最小コストとなる経路上でその直前の格子点の位置が蓄積され、ステップＳ１９で全格子点に対応する最小コストの計算が終わったか否かが判定され、全ての格子点に対する計算が終わるまでステップＳ１７、Ｓ１８の処理が実行され、全ての格子点に対応する計算が終了したと判定されると、ステップＳ２０で終点側から見て最小コストとなる最適経路上で格子点の位置がスタート点まで追跡されて、最適経路が探索される。
【００４９】
ステップＳ１７における最小コスト値の検出においては、図９の横座標ｉ＝ｌ，縦座標ｊ＝ｍの位置の格子点に対して、次式を用いて計算が行なわれる。
【００５０】
【数４】

【００５１】
ステップＳ１８においては、この式に対応して直前の隣接する３つの格子点、すなわち座標（ｌ−１，ｍ），（ｌ−１，ｍ−１），（ｌ，ｍ−１）の格子点の中で、最小コスト値に対応する格子点の位置が蓄積される。
【００５２】
図１３は１つの格子点（Ｕ１，Ｄ１）に対する最小コスト算出の説明図である。この格子点に至るルートとしては、（Ｕ０，Ｄ０）の格子点から上斜め方向に直接に進む経路と、（Ｕ０，Ｄ１）の格子点を経由する経路と、（Ｕ１，Ｄ０）の格子点を経由する経路との３つがある。
【００５３】
これらの経路のうち、途中の経由点までの累積コストｇ（０，１）とｇ（１，０）がすでに計算されているので、それらの値を用いることによってこれらの３つの経路に対応する累積コストを計算することができ、ｇ（１，１）の値は次式によって計算される。
【００５４】
【数５】

【００５５】
従って最小値として１５.９ｄＢが選択される。同様にして全ての格子点に至る全ての経路の計算結果から、コスト最小となる格子点位置をエンド点側からスタート点に向かって逆方向に探索することによって、コスト最小の経路が探索される。
【００５６】
図１４は各格子点に対応する累積コストの計算結果を示す。この図においてエンド点側、すなわち（Ｕ５，Ｄ３）の格子点からコスト最小となる経路上の隣接格子点を探索していくことによって、図１１の結果と同じ経路が探索される。
【００５７】
図１１、または図１４の結果から、図１５のピークマッチング結果が得られる。前述のように周波数の上昇区間側の３つのピークＵ０，Ｕ１、およびＵ２と、下降区間側のピークＤ０とがペアとなっている。これらの中から最適な１対１のペアを選択する必要がある。下降区間側の１つのピークＤ０が正しいものとして、上昇区間側の３つのピークＵ０，Ｕ１、およびＵ２のいずれかがＤ０とペアとして最適であるかを決定するために、ピーク受信電力の対数値の差の絶対値の中で最小のものを選択することにする。これらの絶対値は次式によって計算される。
【００５８】
【数６】

【００５９】
この絶対値の最小の値は１．９ｄＢであり、Ｕ０とＤ０とのペアが最適ペアとして選択される。この絶対値の計算において受信電力の差の計算を真数値で行なうこともできることは当然である。
【００６０】
以上で説明したＤＰマッチングによって決定されたペアに対応するピーク周波数、すなわち図２０で説明したｆupとｆdownとを用いて、以下の式によって、図２の４台の自動車のそれぞれに対する自動車までの距離Ｒと、自動車の速度ｖが計算される。
【００６１】
【数７】

【００６２】
この距離および速度の計算などについては次式の文献に記述されている。
非特許文献１）大久保、藤村、近藤：“60GHz帯自動車用ミリ波レーダ”
ＦＵＪＩＴＳＵ vol.4７，no.4，pp.332-337(07,1995)
以上において本発明のレーダの信号処理装置および信号処理方法について詳細に説明した。本実施形態ではこのような信号処理は図３の１チップマイコン１９によって実行されるものとしたが、この処理をより一般的なコンピュータシステムによって実行することも当然可能である。図１６はそのようなコンピュータシステム、すなわちハードウエア環境の構成ブロック図である。
【００６３】
図１６においてコンピュータシステムは中央処理装置（ＣＰＵ）３０、リードオンリメモリ（ＲＯＭ）３１、ランダムアクセスメモリ（ＲＡＭ）３２、通信インタフェース３３、記憶装置３４、入出力装置３５、可搬型記憶媒体の読み取り装置３６、およびこれらの全てが接続されたバス３７によって構成されている。
【００６４】
記憶装置３４としてはハードディスク、磁気ディスクなど様々な形式の記憶装置を使用することができ、このような記憶装置３４、またはＲＯＭ３１に図４，図８,図１２などのフローチャートに示されたプログラムや、本発明の特許請求の範囲の請求項１１のプログラムなどが格納され、そのようなプログラムがＣＰＵ３０によって実行されることにより、本実施形態における動的計画法を用いたピークマッチング、複数の対象物の識別が可能となる。
【００６５】
このようなプログラムは、プログラム提供者３８側からネットワーク３９、および通信インタフェース３３を介して、例えば記憶装置３４に格納されることも、また市販され、流通している可搬型記憶媒体４０に格納され、読み取り装置３６にセットされて、ＣＰＵ３０によって実行されることも可能である。可搬型記憶媒体４０としてはメモリカード、ＣＤ−ＲＯＭ、フレキシブルディスク、光ディスク、光磁気ディスクなど様々な形式の記憶媒体を使用することができ、このような記憶媒体に格納されたプログラムが読み取り装置３６によって読み取られることにより、本実施形態におけるレーダの信号処理が可能となる。
【００６６】
（付記１） 複数の対象物を検出するレーダにおいて、
レーダからの送信信号と複数の対象物から反射される受信信号とに基づいて、対象物識別用の信号を生成する識別用信号生成手段と、
該識別用信号を受け取り、動的計画法を用いて複数の対象物の識別を行なう対象物識別手段とを備えることを特徴とするレーダの信号処理装置。
【００６７】
（付記２） 前記対象物識別手段が、受信信号のピークに対応して複数の対象物の識別を行なうことを特徴とする付記１記載のレーダの信号処理装置。
（付記３） 前記レーダがＦＭ−ＣＷレーダであり、
前記対象物識別手段が、ＦＭ周波数の上昇区間と下降区間とにおける受信信号のピークのマッチングによって、前記複数の対象物の識別を行なうことを特徴とする付記２記載のレーダの信号処理装置。
【００６８】
（付記４） 前記ピークマッチングにおいて、送信信号と受信信号とから得られるビート信号を前記識別用信号として、該ビート信号のフーリエ変換の結果として得られる受信電力のピークを用いることを特徴とする付記３記載のレーダの信号処理装置。
【００６９】
（付記５） 前記ピークマッチングにおいて、受信電力の周波数に対するピークを、前記上昇区間と下降区間との間で対応させることを特徴とする付記４記載のレーダの信号処理装置。
【００７０】
（付記６） 前記ピークマッチングにおいて、前記上昇区間におけるピークと下降区間におけるピークとのマッチングのコストとして、ピーク受信電力の差の絶対値を用いることを特徴とする付記５記載のレーダの信号処理装置。
【００７１】
（付記７） 前記ピークマッチングにおいて、前記上昇区間におけるピークと下降区間におけるピークとのマッチングのコストとして、ピーク受信電力の対数値の差の絶対値を用いることを特徴とする付記５記載のレーダの信号処理装置。
【００７２】
（付記８） 前記ピークマッチングにおいて、前記上昇区間と下降区間との間で１つのピークと複数のピークとが対応する結果が得られた時、該複数のピークのうちで該１つのピークとの受信電力の差の絶対値が最小となるピークを該１つのピークに対するマッチングの結果とすることを特徴とする付記５記載のレーダの信号処理装置。
【００７３】
（付記９） 前記ピークマッチングにおいて、前記上昇区間と下降区間との間で１つのピークと複数のピークが対応する結果が得られた時、該複数のピークのうちで該１つのピークとの受信電力の対数値の差の絶対値が最小となるピークを該１つのピークに対するマッチングの結果とすることを特徴とする付記５記載のレーダの信号処理装置。
【００７４】
（付記１０） 複数の対象物を検出するレーダの信号処理方法において、
レーダからの送信信号と複数の対象物から反射される受信信号とに基づいて、対象物識別用の信号を生成し、
該識別用信号を使用して、動的計画法によって複数の対象物の識別を行なうことを特徴とするレーダの信号処理方法。
【００７５】
（付記１１） 複数の対象物を検出するレーダの信号処理を行なう計算機によって使用されるプログラムにおいて、
レーダからの送信信号と複数の対象物から反射される受信信号とに基づいて、対象物識別用の信号を生成する手順と、
該識別用信号を使用して、動的計画法によって複数の対象物の識別を行なう手順とを計算機に実行させるためのプログラム。
【００７６】
（付記１２） 複数の対象物を検出するレーダの信号処理を行なう計算機によって使用される記憶媒体において、
レーダからの送信信号と複数の対象物から反射される受信信号とに基づいて、対象物識別用の信号を生成するステップと、
該識別用信号を使用して、動的計画法によって複数の対象物の識別を行なうステップとを計算機に実行させるためのプログラムを格納した計算機読み出し可能可搬型記憶媒体。
【００７７】
【発明の効果】
以上詳細に説明したように、本発明によればノイズやマルチパスの影響がある場合にも、動的計画法を用いてピークマッチングを行なうことによって誤ったペアリングの可能性を減少させることができ、複数の計測対象物を識別するためのレーダの実用性の向上に寄与するところが大きい。
【図面の簡単な説明】
【図１】本発明の信号処理装置の原理構成ブロック図である。
【図２】本実施形態におけるレーダ使用状況の説明図である。
【図３】本実施形態におけるＦＭ−ＣＷレーダの構成ブロック図である。
【図４】計測対象物検出処理の全体フローチャートである。
【図５】周波数上昇区間におけるピークの例である。
【図６】周波数下降区間におけるピークの例である。
【図７】ＤＰマッチングにおける重みを説明する図である。
【図８】ＤＰマッチングの第１の方法に対応する処理のフローチャートである。
【図９】図７に対応する格子点の座標の説明図である。
【図１０】第１の方法における最小コスト値算出例の説明図である。
【図１１】第１の方法におけるＤＰマッチング結果としての最適経路を示す図である。
【図１２】ＤＰマッチングの第２の方法に対応する処理のフローチャートである。
【図１３】第２の方法における格子点までの累積コストの計算例を説明する図である。
【図１４】第２の方法における各格子点までの経路の累積コストの計算結果を示す図である。
【図１５】ＤＰマッチングの結果に対する最適ピークマッチングを説明する図である。
【図１６】本実施形態におけるプログラムのコンピュータへのローディングの説明図である。
【図１７】ＦＭ−ＣＷレーダの従来例の基本構成ブロック図である。
【図１８】ＦＭ−ＣＷレーダにおける送信波と受信波の周波数の説明図である。
【図１９】ＦＭ−ＣＷレーダによる静止している対象物に対する計測法を説明する図である。
【図２０】ＦＭ−ＣＷレーダによる移動している対象物の計測方法を説明する図である。
【符号の説明】
１ 信号処理装置
２ 識別用信号生成手段
３ 対象物識別手段
１０ 同期信号発生器
１１ 三角波発生器
１２ 電圧制御発振器（ＶＣ０）
１３ 送信（Ｔ）アンテナ
１５ 受信（Ｒ）アンテナ
１６ ミキサ
１７ アンプ
１８ Ａ／Ｄコンバータ
１９ １チップマイコン[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a radar signal processing system, and more particularly to a radar signal processing apparatus and a processing method for accurately detecting a plurality of objects in an FM-CW radar using millimeter waves.
[0002]
[Prior art]
Radio wave radar using millimeter waves or the like has begun to be used as a vehicle cruise control sensor. As a millimeter wave radar system, an FM-CW (Frequency Modulated Continuous Wave) system has become the mainstream.
[0003]
The FM-CW radar performs frequency modulation (FM) on a continuous wave, that is, a CW signal, and uses it as a transmission signal. 17 to 20 are explanatory diagrams of the operation of this FM-CW radar.
[0004]
FIG. 17 is a block diagram showing a conventional FM-CW radar. In the figure, the radio wave transmitted from the millimeter wave transmitter 100 is reflected by the measurement object 101, and the reflected wave is received by the millimeter wave receiver 102. The transmission signal and the reception signal are given to the beat frequency generation unit 103, the beat signal is given to the signal processing unit 104, and the distance from the radio wave flight time and Doppler shift to the measurement target 101 and the moving speed of the measurement target are determined. Desired.
[0005]
FIG. 18 is an explanatory diagram of the frequency modulation operation in the FM-CW radar. In the figure, a transmission wave indicated by a solid line is modulated and transmitted in a form that repeats an ascending interval in which the frequency f rises linearly and a descending interval in which the frequency f falls linearly. The difference between the maximum value and the minimum value of the frequency is the frequency modulation width, and the average value of the maximum value and the minimum value is the center frequency. The received wave indicated by the dotted line repeats the frequency change in the same manner as the transmitted wave, but the time of flight of the radio wave is obtained by comparing the times when both waves are maximum, for example.
[0006]
FIG. 19 is an explanatory diagram of beat signal generation by the beat frequency generation unit 103 when detecting a stationary measurement object. When the object is stationary, the frequency difference fr between the transmitted wave and the received wave is the same and constant in the rising and falling intervals, as in FIG. The received power W as a result of signal processing, that is, the result of FFT (Fast Fourier Transform), has a peak in both the rising and falling intervals with respect to the frequency difference fr.
[0007]
FIG. 20 is an explanatory diagram of the operation when the measurement object is moving. When the measurement object is moving, the difference in frequency between the ascending section and the descending section differs depending on the Doppler shift fd. When fr in FIG. 19 is used, the difference is fr−fd in the ascending section and fr + fd in the descending section. Thus, the received power value as a result of the FFT has a peak at fup in the rising section and fdown in the falling section. Here, the relationship between these frequencies is as follows.
[0008]
fup = fr−fd
fdown = fr + fd
As a conventional technique related to such FM-CW radar, there is the following document.
[0009]
[Patent Document 1]
Japanese Patent Laid-Open No. 9-152477 “FMCW Radar Device”
[Patent Document 2]
JP 2000-257927 “Signal processing method of millimeter wave sensor”
[0010]
In Patent Document 1, in an FM-CW radar apparatus having two reception channels, a phase difference between peak frequency components having the same frequency between the two reception channels is calculated for each of the frequency rising and falling intervals. When the absolute value of the difference in phase difference is smaller than a predetermined value, the distance to the object and the relative velocity can be obtained even when there are multiple objects by matching (pairing) the peak frequency components. A radar device that can be detected is disclosed.
[0011]
Patent Document 2 discloses a signal processing method of a millimeter wave sensor that can prevent erroneous pairing that occurs due to the influence of rain, fog, or the like in peak matching by using long-time background data and short-time background data. It is disclosed.
[0012]
[Problems to be solved by the invention]
As described above, in the peak matching method of the FM-CW radar, the Fourier transform is separately performed in the rising and falling intervals of the FM-modulated frequency, and the pair of the peaks having the reception power close between the detected peaks is performed. It is. However, actually, the detected peak may include, for example, a peak due to a reflected signal from a guardrail or a road sign, noise, or the like. Therefore, there are cases where the number of peaks differs between the ascending section and the descending section, and the method of Patent Document 1 has a problem that a peak due to noise and a peak due to an object are erroneously paired.
[0013]
In Patent Document 2, a method is employed in which the transmission direction of the radar is changed and erroneous pairing is suppressed using received data corresponding to a plurality of directions. However, this method has a problem that erroneous pairing cannot be suppressed when the direction of the radar is fixed in one direction.
[0014]
In view of the above problems, an object of the present invention is to provide a radar signal processing device and a processing method capable of correctly executing peak matching corresponding to a plurality of objects even when there is an influence of noise or the like. It is to be.
[0015]
[Means for Solving the Problems]
FIG. 1 is a block diagram showing the principle configuration of a radar signal processing apparatus according to the present invention. In FIG. 1, the signal processing apparatus 1 includes an identification signal generation unit 2 and an object identification unit 3.
[0016]
The identification signal generating means 2 generates a signal for identifying an object, for example, a beat signal, based on reception signals reflected from a plurality of objects and transmission signals from the radar. The object identification unit 3 receives a signal output from the identification signal generation unit 2 and identifies a plurality of objects using dynamic programming, that is, dynamic programming.
[0017]
In the embodiment of the invention, the object identifying means 3 can also identify a plurality of objects corresponding to the peak of the received signal.
In the embodiment, the radar is an FM-CW radar, and the object identifying means 3 can identify a plurality of objects by peak matching of received signals in the rising and falling sections of the FM frequency. In the case of peak matching, the beat signal obtained from the transmission signal and the reception signal can be used as a recognition signal, and the peak of the reception power obtained as a result of the Fourier transform of the signal can also be used. And the descending section can be made to correspond.
[0018]
In the embodiment, the absolute value of the difference in peak received power may be used as the cost for matching the peak in the rising section and the peak in the falling section in peak matching, or the absolute value of the logarithmic difference in peak received power. Can also be used.
[0019]
Furthermore, in the embodiment, when a result corresponding to one peak and a plurality of peaks is obtained in the peak matching between the rising section and the falling section, reception of one of the plurality of peaks is received. The peak with the smallest absolute value of the power difference can be used as a result of matching with the one peak, or the absolute value of the difference in the logarithmic value of the received power with one of the peaks can be minimized. Can be used as a result of matching for the one peak.
[0020]
In the radar signal processing method of the present invention, a signal for identifying an object is generated based on a reception signal reflected from a plurality of objects and a transmission signal from the radar, and a motion is generated using the identification signal. A method of identifying a plurality of objects is used by the genetic programming method.
[0021]
As the program and storage medium of the present invention, a program for realizing the signal processing method and a computer-readable portable storage medium may be used.
[0022]
As described above, according to the present invention, a plurality of objects are identified using dynamic programming.
[0023]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 2 is an explanatory diagram of the usage status of the radar in the embodiment of the present invention. In the figure, a beam is emitted from the radar of the car that is running the farthest, and in order to perform cruise control, the distance between the four cars that are running in front and the speed of each car are measured. The situation where is being done is shown.
[0024]
FIG. 3 is a configuration block diagram of the FM-CW radar in the present embodiment. In the figure, the output of the synchronization signal (Sync) generator 10 is given to the triangular wave oscillator 11, and the voltage-controlled oscillator (VCO) 12 converts the frequency-modulated transmission wave as described in FIG. 18 into a transmission (T) antenna. 13 to the measurement object side.
[0025]
The reception signal received by the reception (R) antenna 15 is given to the mixer 16 corresponding to the beat frequency generation unit 103 in FIG. 17 together with the output of the voltage controlled oscillator 12, and the beat signal is output and amplified by the amplifier 17. The A / D converter 18 converts it into a digital signal in a form synchronized with the output of the synchronization signal generator 10, and the one-chip microcomputer 19 performs data processing such as Fourier transform and peak matching (pairing) to identify the plural The distance to the measurement object and the movement distance of the object are output to the outside.
[0026]
FIG. 4 is an overall flowchart of the object detection process in the present embodiment. In the figure, when processing is started, frequency modulation, that is, digitization of the beat signal synchronized with the period of FM is first performed in step S1, and the beat signal is divided into frequency rising and falling intervals in step S2. Fourier transform is performed, and threshold processing and peak detection are performed in step S3. A method for setting the threshold will be described later.
[0027]
In step S4, peak matching (pairing) is performed by a dynamic programming (dynamic programming, DP) method. At this time, for example, when a plurality of peaks in the rising section and a single peak in the falling section are matched (paired), an optimum peak match is set to make this matching one-to-one in step S5. In step S6, the distance to the object and the moving speed of the object are finally converted from the peak frequency and calculated, and the process ends.
[0028]
5 and 6 show the received power as a result of the fast Fourier transform (FFT) of the digitized beat signal with respect to the frequency. FIG. 5 shows the frequency increase period, and FIG. 6 shows the decrease. The result in the section is shown.
[0029]
In the rising section of FIG. 5, six peaks larger than the threshold are detected, whereas in the falling section of FIG. 6, four peaks are detected. Assuming that four cars are traveling ahead as described in FIG. 2, two of the six peaks in FIG. 5 are considered to be noise caused by guardrails, road signs, etc., or peaks caused by multipath. . U1 and U2 are estimated to be such peaks from the strength of the received power, but in this embodiment, peak matching including such peaks is performed.
[0030]
Here, for example, there are two methods for determining the threshold value. The first method is a method using a predicted value of received power reflected from an object. For example, the radar effective reflection cross section (RCS) for passenger cars etc. is 1 m in cross section. ² If the value reflected from the sphere is a reference value, that is, 0 dBsm, the value is about 10 dBsm, for example. The received power of the radar can be predicted from the RCS of the object, the approximate distance, the radar constants, etc., and the threshold is determined as a value including the margin and lower than the received power predicted value, for example.
[0031]
The second method is a determination method using noise. The received power of noise such as thermal noise can be predicted from the radar components. A threshold value can be determined by including a margin. Furthermore, the threshold value can be determined by combining the first method and the second method.
[0032]
Next, peak matching in step S4 in FIG. 4 will be described. FIG. 7 is an example of setting weights in DP matching. In the figure, U0 to U5 indicating the six peaks in the frequency increasing section of FIG. 5 in the horizontal direction are arranged in order of increasing frequency, and D0 indicating the four peaks in the decreasing section of FIG. 6 in the vertical direction. ˜D3 are arranged in order of increasing frequency. Here, it is assumed that the order of the peak frequencies corresponds to the order of the four automobiles in FIG.
[0033]
In the DP method, a start point and an end point are set as appropriate, and a route that minimizes the cost from the start point to the end point is searched. In the lattice-like arrangement of FIG. Peak matching is performed corresponding to the points.
[0034]
In FIG. 7, it is assumed that the end point is reached by basically tracing in an upward direction from the start point to the end point. Therefore, inside the grid, a path that advances only in the upward diagonal direction, the upward direction, or the lateral (right) direction is set, and when going from one grid point to another grid point in the upward diagonal direction, the weight is set to 1, and the upward direction Alternatively, the weight is set to 2 when proceeding in the horizontal direction.
[0035]
For example, when a part of the optimum path is set from the lattice point (U0, D0) to the lattice point (U1, D0) in the horizontal direction, the two peaks U0, U1 in FIG. This means that a pair with one peak D0 is obtained as a matching result, and thus there is less possibility of obtaining such a correspondence for a correspondence where one peak corresponds to a plurality of peaks. Therefore, by increasing the weight when calculating the cost, the upward diagonal direction is easily selected as the direction of the optimum route.
[0036]
FIG. 8 is a process flowchart of the first method of DP matching. In the first method, after calculating the cost from the start point of FIG. 7 to the grid point (U0, D0) as an initial cost, the cost of proceeding to which adjacent grid point among the adjacent grid points of this grid point is calculated. The minimum cost value is detected as a minimum cost value, and the adjacent grid points in that direction are determined as a part of the optimum route. Next, the optimal path is searched by repeating the process of detecting which direction of the three adjacent grid points in the upward diagonal direction, the upward direction, or the horizontal direction from the adjacent grid point is the smallest. become.
[0037]
When the process is started in FIG. 8, the initial cost value is calculated in step S11. Here, it is assumed that this cost value is obtained as an absolute value of the difference between the logarithmic values of the peak received power. In FIG. 7, since the peak power value is expressed in dB, the initial cost value is calculated by subtracting the peak value corresponding to D0 from the peak value corresponding to U0. With this value as d (0, 0), the value of the initial cost g is obtained by the following equation.
[0038]
[Expression 1]

[0039]
FIG. 9 is an explanatory diagram of variables i and j used in this equation. In the figure, i corresponds to the value of the coordinate on the horizontal axis corresponding to FIG. 7, j corresponds to the value of the coordinate corresponding to the vertical axis, and the value of i corresponds to values 0 to 5, corresponding to FIG. Is in the range of 0 to 3.
[0040]
In step S12 in FIG. 8, the minimum cost value is detected and stored in a path from a certain lattice point to an adjacent lattice point. In step S13, the position information of the adjacent grid point that is the minimum cost is accumulated, and the processing in and after step S12 is repeated until it is determined in step S14 that the end point position has been reached, and when the end point position is reached, DP matching processing is performed. finish.
[0041]
In the minimum cost value detection process in step S12, calculation is performed using the following equation.
[0042]
[Expression 2]

[0043]
Here, k starts from 0 and corresponds to the number of repetitions of steps S12 to S14, and F is the weight in each direction described with reference to FIG. d is the absolute value of the difference between the logarithmic values of the peak received power, and g is one grid when calculating which of the costs from one grid point to three adjacent grid points is minimized. This is the cost value corresponding to the searched route to the point. The smallest one of the three values is detected, and the position of the adjacent grid point corresponding to the minimum cost value is substituted into i (k + 1) and j (k + 1) indicating the coordinates, and the position is stored in step S13. Is done.
[0044]
FIG. 10 is an explanatory diagram of the minimum cost value detection process for determining which direction of the three adjacent grid points to proceed from the lower left grid point (U0, D0). In this process, the value of g when traveling in the upward direction, the diagonally upward direction, and the lateral direction is calculated by the following equation.
[0045]
[Equation 3]

[0046]
Of these three, the cost g (1,0) proceeding in the horizontal direction is minimized, so that the path traveling in the horizontal direction is determined as a part of the optimum path as shown in FIG.
By repeating the processes in steps S12 and S13 in FIG. 8, the optimum paths ahead are determined one after another from the state in FIG. FIG. 11 shows the result. From this result, the three peaks U0 to U2 in the rising section of FIG. 5 correspond to one peak D0 in the falling section of FIG. 6, and the three peaks U3 to U5 in the rising section are the three peaks D1 in the falling section. A matching result corresponding to 1 to 1 is obtained for each of .about.D3.
[0047]
FIG. 12 is a process flowchart of the second method of DP matching. In this second method, for each of the lattice points in FIG. 7, that is, each of the 24 lattice points, the cost of the route having the smallest cost among the routes from the start point to the lattice point is reduced. After detecting the minimum cost value of each grid position, the position of the grid point having the minimum cost is traced from the end point side to reach the start point, and the optimum path is searched.
[0048]
When the process is started in FIG. 12, first, in step S16, the cost value for reaching g (0,0), that is, the grid point of (U0, D0) from the start point is calculated in the same manner as in step S11 of FIG. After that, in step S17, the minimum cost value corresponding to each grid point is detected and stored as described above, and the path having the minimum cost when viewed from the position of the grid point calculated in step S18. The position of the immediately preceding grid point is accumulated above, and it is determined in step S19 whether or not the calculation of the minimum cost corresponding to all the grid points has been completed. Is executed, and it is determined that the calculation corresponding to all the lattice points has been completed, the position of the lattice point is scanned on the optimum path that has the minimum cost when viewed from the end point side in step S20. Is tracked until over preparative point, the optimal path is searched.
[0049]
In the detection of the minimum cost value in step S17, calculation is performed using the following equation for the lattice point at the position of the abscissa i = 1 and the ordinate j = m in FIG.
[0050]
[Expression 4]

[0051]
In step S18, the immediately preceding three adjacent lattice points corresponding to this expression, that is, the lattice points at coordinates (l-1, m), (l-1, m-1), (l, m-1). The position of the grid point corresponding to the minimum cost value is stored.
[0052]
FIG. 13 is an explanatory diagram of the minimum cost calculation for one grid point (U1, D1). The route to this lattice point includes a route that goes directly diagonally upward from the lattice point of (U0, D0), a route that passes through the lattice point of (U0, D1), and a lattice point of (U1, D0). There are three routes:
[0053]
Of these routes, the accumulated costs g (0, 1) and g (1, 0) to the waypoints in the middle have already been calculated, so these values are used to correspond to these three routes. The accumulated cost can be calculated, and the value of g (1,1) is calculated by the following equation.
[0054]
[Equation 5]

[0055]
Therefore, 15.9 dB is selected as the minimum value. Similarly, from the calculation results of all the routes to all the lattice points, the least cost route is searched by searching in the reverse direction from the end point side to the start point for the lattice point position that minimizes the cost. .
[0056]
FIG. 14 shows the calculation result of the accumulated cost corresponding to each grid point. In this figure, the same path as the result of FIG. 11 is searched by searching for the adjacent grid point on the path having the minimum cost from the end point side, that is, the grid point of (U5, D3).
[0057]
From the result of FIG. 11 or FIG. 14, the peak matching result of FIG. 15 is obtained. As described above, the three peaks U0, U1, and U2 on the frequency rising section side and the peak D0 on the falling section side are paired. It is necessary to select an optimal one-to-one pair from these. Assuming that one peak D0 on the descending section side is correct, in order to determine which of the three peaks U0, U1 and U2 on the rising section side is optimal as a pair with D0, the logarithmic value of the peak received power The smallest one of the absolute values of the differences is selected. These absolute values are calculated by the following equation.
[0058]
[Formula 6]

[0059]
The minimum absolute value is 1.9 dB, and a pair of U0 and D0 is selected as the optimal pair. In the calculation of the absolute value, it is natural that the difference in received power can be calculated with a true value.
[0060]
Using the peak frequency corresponding to the pair determined by the DP matching described above, that is, fup and fdown described in FIG. 20, the distance to the vehicle with respect to each of the four vehicles in FIG. R and the vehicle speed v are calculated.
[0061]
[Expression 7]

[0062]
The calculation of distance and speed is described in the following equation.
Non-patent document 1) Okubo, Fujimura, Kondo: "60GHz band automotive millimeter wave radar"
FUJITSU vol.47, no.4, pp.332-337 (07,1995)
The radar signal processing apparatus and signal processing method of the present invention have been described in detail above. In the present embodiment, such signal processing is executed by the one-chip microcomputer 19 in FIG. 3, but it is naturally possible to execute this processing by a more general computer system. FIG. 16 is a block diagram showing the configuration of such a computer system, that is, a hardware environment.
[0063]
In FIG. 16, the computer system is a central processing unit (CPU) 30, a read only memory (ROM) 31, a random access memory (RAM) 32, a communication interface 33, a storage device 34, an input / output device 35, and a portable storage medium reading device. 36, and a bus 37 to which all of these are connected.
[0064]
As the storage device 34, various types of storage devices such as a hard disk and a magnetic disk can be used. Such a storage device 34 or a ROM 31 has programs shown in the flowcharts of FIGS. 4, 8, and 12. The program of claim 11 of the claims of the present invention is stored, and when such a program is executed by the CPU 30, peak matching using the dynamic programming method in the present embodiment, a plurality of objects Can be identified.
[0065]
Such a program is stored in, for example, the storage device 34 from the program provider 38 side via the network 39 and the communication interface 33, or stored in a portable storage medium 40 that is commercially available and distributed. It can also be set in the reading device 36 and executed by the CPU 30. As the portable storage medium 40, various types of storage media such as a memory card, a CD-ROM, a flexible disk, an optical disk, and a magneto-optical disk can be used, and a program stored in such a storage medium is read by the reading device 36. Thus, radar signal processing in the present embodiment is possible.
[0066]
(Supplementary note 1) In a radar that detects a plurality of objects,
Identification signal generating means for generating a signal for identifying an object based on a transmission signal from a radar and reception signals reflected from a plurality of objects;
A radar signal processing apparatus comprising: an object identification unit that receives the identification signal and identifies a plurality of objects using dynamic programming.
[0067]
(Supplementary note 2) The radar signal processing apparatus according to supplementary note 1, wherein the target object identification means identifies a plurality of target objects corresponding to a peak of a received signal.
(Appendix 3) The radar is FM-CW radar,
3. The radar signal processing apparatus according to appendix 2, wherein the object identifying means identifies the plurality of objects by matching received signal peaks in an FM frequency rising section and a falling section.
[0068]
(Additional remark 4) In the said peak matching, the peak of the received power obtained as a result of the Fourier-transform of this beat signal is used for the said beat signal obtained from a transmission signal and a received signal as said identification signal. 3. The radar signal processing apparatus according to 3.
[0069]
(Supplementary note 5) The radar signal processing device according to supplementary note 4, wherein, in the peak matching, a peak with respect to a frequency of received power is caused to correspond between the rising section and the falling section.
[0070]
(Additional remark 6) In the said peak matching, the absolute value of the difference of peak received power is used as the cost of matching with the peak in the said rising area, and the peak in a falling area, The signal processing apparatus of the additional mark 5 characterized by the above-mentioned .
[0071]
(Supplementary note 7) The radar according to supplementary note 5, wherein, in the peak matching, an absolute value of a difference between logarithmic values of peak received power is used as a matching cost between a peak in the rising section and a peak in the falling section. Signal processing device.
[0072]
(Supplementary Note 8) In the peak matching, when a result corresponding to one peak and a plurality of peaks is obtained between the ascending section and the descending section, the one peak among the plurality of peaks The radar signal processing apparatus according to appendix 5, wherein a peak having a minimum absolute value of a difference in received power is set as a result of matching with the one peak.
[0073]
(Supplementary Note 9) In the peak matching, when a result corresponding to one peak and a plurality of peaks is obtained between the rising section and the falling section, reception of the one peak among the plurality of peaks is received. 6. The radar signal processing apparatus according to appendix 5, wherein a peak having a minimum absolute value of a difference between logarithmic values of power is set as a result of matching with the one peak.
[0074]
(Supplementary Note 10) In a radar signal processing method for detecting a plurality of objects,
Based on the transmission signal from the radar and the reception signals reflected from a plurality of objects, a signal for object identification is generated,
A radar signal processing method, wherein a plurality of objects are identified by dynamic programming using the identification signal.
[0075]
(Supplementary Note 11) In a program used by a computer that performs radar signal processing to detect a plurality of objects,
A procedure for generating a signal for object identification based on a transmission signal from a radar and reception signals reflected from a plurality of objects;
A program for causing a computer to execute a procedure for identifying a plurality of objects by dynamic programming using the identification signal.
[0076]
(Supplementary Note 12) In a storage medium used by a computer that performs radar signal processing to detect a plurality of objects,
Generating a signal for object identification based on a transmission signal from a radar and reception signals reflected from a plurality of objects;
A computer-readable portable storage medium storing a program for causing a computer to execute a step of identifying a plurality of objects by dynamic programming using the identification signal.
[0077]
【The invention's effect】
As described above in detail, according to the present invention, the possibility of erroneous pairing can be reduced by performing peak matching using dynamic programming even when there is an influence of noise or multipath. It is possible to greatly improve the practicality of radar for identifying a plurality of measurement objects.
[Brief description of the drawings]
FIG. 1 is a block diagram showing the principle configuration of a signal processing apparatus according to the present invention.
FIG. 2 is an explanatory diagram of a radar usage situation in the present embodiment.
FIG. 3 is a configuration block diagram of an FM-CW radar in the present embodiment.
FIG. 4 is an overall flowchart of measurement object detection processing;
FIG. 5 is an example of a peak in a frequency increase section.
FIG. 6 is an example of a peak in a frequency decreasing section.
FIG. 7 is a diagram illustrating weights in DP matching.
FIG. 8 is a flowchart of processing corresponding to a first method of DP matching.
FIG. 9 is an explanatory diagram of coordinates of lattice points corresponding to FIG. 7;
FIG. 10 is an explanatory diagram of a minimum cost value calculation example in the first method.
FIG. 11 is a diagram illustrating an optimum path as a DP matching result in the first method.
FIG. 12 is a flowchart of processing corresponding to a second method of DP matching.
FIG. 13 is a diagram illustrating an example of calculating an accumulated cost up to a lattice point in the second method.
FIG. 14 is a diagram illustrating a calculation result of a cumulative cost of a route to each lattice point in the second method.
FIG. 15 is a diagram for explaining optimum peak matching with respect to a DP matching result;
FIG. 16 is an explanatory diagram of loading of a program into a computer according to the present embodiment.
FIG. 17 is a basic configuration block diagram of a conventional example of an FM-CW radar.
FIG. 18 is an explanatory diagram of frequencies of transmission waves and reception waves in the FM-CW radar.
FIG. 19 is a diagram for explaining a measurement method for a stationary object by FM-CW radar.
FIG. 20 is a diagram for explaining a method of measuring a moving object by FM-CW radar.
[Explanation of symbols]
1 Signal processing equipment
2 Identification signal generation means
3 Object identification means
10 Sync signal generator
11 Triangular wave generator
12 Voltage controlled oscillator (VC0)
13 Transmitting (T) antenna
15 Receiving (R) antenna
16 Mixer
17 Amplifier
18 A / D converter
19 One-chip microcomputer

Claims (5)

In FM-CW radar that detects multiple objects,
Identification signal generating means for generating a signal for identifying an object based on a transmission signal from the FM-CW radar and reception signals reflected from a plurality of objects;
Object identification means for receiving the identification signal and identifying a plurality of objects by performing pairing using dynamic programming with the absolute value of the difference in peak received power as a cost ;
A signal processing apparatus for FM-CW radar, comprising: means for selecting a minimum one of absolute values of differences in peak received power when one to a plurality of pairs are formed . 2. The FM-CW radar according to claim 1, wherein the object identification means identifies the plurality of objects by performing pairing of received signal peaks during an increase period and a decrease period of an FM frequency. Signal processing equipment. In an FM-CW radar signal processing method for detecting a plurality of objects,
Based on a transmission signal from the FM-CW radar and reception signals reflected from a plurality of objects, a signal for object identification is generated,
Using the identification signal, the row stomach identification of a plurality of objects by performing pairing by dynamic programming to cost the absolute value of the difference between the peak received power,
A signal processing method for an FM-CW radar, wherein a minimum one of absolute values of differences in peak received power is selected when one to a plurality of pairs are formed . In a program used by a computer that performs signal processing of an FM-CW radar that detects a plurality of objects,
A procedure for generating a signal for object identification based on a transmission signal from the FM-CW radar and reception signals reflected from a plurality of objects;
Using the identification signal, a procedure for identifying a plurality of objects by performing pairing according to dynamic programming using the absolute value of the difference in peak received power as a cost ;
A program for causing a computer to execute a procedure for selecting a minimum one of absolute values of differences in peak received power when a pair of one to a plurality of pairs is made . In a storage medium used by a computer that performs signal processing of FM-CW radar that detects a plurality of objects,
Generating a signal for object identification based on a transmission signal from the FM-CW radar and a reception signal reflected from a plurality of objects;
Using the identification signal to identify a plurality of objects by pairing by dynamic programming with the absolute value of the difference in peak received power as a cost ;
A computer-readable portable storage medium storing a program for causing a computer to execute a step of selecting a minimum one of absolute values of differences in peak received power when a pair of one to a plurality of pairs is made.

Images