1、( )( )()at nTag ng tgnT( )g n( )ag t数字滤波器的阶跃响应 模仿模拟滤波器的阶跃响应T 抽样周期( )( )h nH z( )1u nzz ( )( )* ( )( )( )1g nu nh nzG zH zz( )( )aah tHs( )1u ts( )( )*( )1( )( )aaaag tu th tG sHss1( )( )aag tLG s( )( )G zZT g n1( )( )( )aaaHsG sHss( )()ag ngnT1( )( )zH zG zz1( )( )aaG sHss( )( )1zG zH zz( )( )()at n
2、Tag ng tgnT( )( )sTaz eG zG s12akGsjkTT( )211sTkzaeGsjzH zzkTT212aksjkTTsjkHT阶跃响应不变法同样有频率响应的混叠失真现象但比冲激响应不变法要小。使数字滤波器的频率响应与模拟滤波器的频率响应相似。冲激响应不变法、阶跃响应不变法:时域模仿逼近缺点是产生频率响应的混叠失真:, 1:,T T 12Ttg 1s Tze11sin2cos2TT1111s Ts Tee1111zsz 11szs12Ttg 1111222222TTjjTTjjeejee11112222TTjjTTjjeesjee 11112222s Ts Ts Ts
3、 Teeee1111zz1s Tzesj 11sj 12s Te12Tc tg 1111zsczcszcs12Tc tg 2cTcc 122cccTc tgc tg 2cccctg 1 1 12Tc2222()()czc01z1111112jjzesccjc tgjze cscjzcscj 01z01z2c tg 00 00 11/T 1112 tgc分段常数型模拟滤波器经变换后仍为分段常数型数字滤波器,但临界频率点产生畸变1112c tg 按 设计模拟滤波器,经双线性变换后,即可得到 为截止频率的数字滤波器111111111( )( )1aazs czzH zHsHcz可分解成级联的低阶子系统可分解成并联的低阶子系统1111( )( ) 1,2,.,iiazs czH zHsim其中:12( )( )( )( )maaaaHsHs HsHs12( )( )( )( )mH zH z HzHz12( )( )( )( )maaaaHsHsHsHs12( )( )( )( )mH zH zHzHz1111( )( ) 1,2,.,iiazs czH zHsim其中: