公开(公告)号：US20100217733A1

公开(公告)日：2010-08-26

申请号：US12680678

申请日：2008-09-30

申请人： Shingo Shimoda ,  Hidenori Kimura ,  Masaaki Yamaoka ,  Hideki Kajima

发明人： Shingo Shimoda ,  Hidenori Kimura ,  Masaaki Yamaoka ,  Hideki Kajima

IPC分类号： G06N3/063

CPC分类号： G06N3/063 ,  G05B2219/39031 ,  G06N3/04

摘要： In order to utilize variable neuron thresholds and extended Hebb's rule in a neural network for proper control, a neuron device (101) for simulating a nerve cell comprises a threshold storage unit (102) storing a threshold variable θ and two threshold coefficients Δθ1 and Δθ2; an input reception unit (103) receiving one or multiple input signal values at predetermined time intervals; an output unit (104) outputting, as an output signal value, a value “1” for indicating that the neuron device (101) is firing when the sum total s of received input signal values is equal to or greater than the value of the stored threshold variable θ, and otherwise a value “0” for indicating that the neuron device (101) is resting; and a threshold updating unit (105) calculating Δθ1X+Δθ2(X−1) using the output signal value X and the stored threshold coefficients Δθ1 and Δθ2 and updating the value of the threshold variable θ stored in the threshold storage unit (102) by increasing it by the calculation result.

摘要翻译： 为了在神经网络中利用可变神经元阈值和扩展Hebb规则进行适当的控制，用于模拟神经细胞的神经元装置（101）包括存储阈值变量的阈值存储单元（102） 和两个阈值系数＆Dgr;＆thetas; 1和＆Dgr;＆thetas; 2; 输入接收单元，以预定的时间间隔接收一个或多个输入信号值; 输出单元（104）当接收的输入信号值的总和等于或大于该值时输出用于指示神经元装置（101）正在点火的值“1”作为输出信号值 存储的阈值变量;以及否则指示神经元设备（101）休息的值“0” 和阈值更新单元（105），使用输出信号值X和所存储的阈值系数＆Dgr;＆amp; 1和＆Dgr;＆amp; thetas; 2和更新所述阈值更新单元（105）来计算＆Dgr;＆Theta; 1X +＆Dgr; 阈值变量的值; 通过计算结果将其存储在阈值存储单元（102）中。

公开(公告)号：US08620844B2

公开(公告)日：2013-12-31

申请号：US12680678

申请日：2008-09-30

申请人： Shingo Shimoda ,  Hidenori Kimura ,  Masaaki Yamaoka ,  Hideki Kajima

发明人： Shingo Shimoda ,  Hidenori Kimura ,  Masaaki Yamaoka ,  Hideki Kajima

IPC分类号： G05B13/02

CPC分类号： G06N3/063 ,  G05B2219/39031 ,  G06N3/04

摘要： Using variable neuron thresholds and extended Hebb's rule in a neural network, a neuron device for simulating a nerve cell includes a threshold storage unit storing a threshold variable θ and threshold coefficients Δθ1 and Δθ2; an input reception unit receiving one or more input signal values at predetermined time intervals; an output unit outputting an output signal value “1” indicating that the neuron device is firing when the sum total s of received input signal values is equal to or greater than the value of the stored threshold variable θ, or a value “0” indicating that the neuron device is resting; and a threshold updating unit calculating Δθ1X+Δθ2(X−1) using the output signal value X and the stored threshold coefficients Δθ1 and Δθ2 and updating the value of the threshold variable θ stored in the threshold storage unit by increasing it by the calculation result.

摘要翻译： 在神经网络中使用可变神经元阈值和扩展Hebb规则，用于模拟神经细胞的神经元装置包括存储阈值变量θ和阈值系数Deltatheta1和Deltatheta2的阈值存储单元; 输入接收单元，以预定的时间间隔接收一个或多个输入信号值; 当所接收的输入信号值的总和为等于或大于所存储的阈值变量theta的值时，输出表示神经元装置正在点火的输出信号值“1”的输出单元，或表示 神经元器件休息; 以及阈值更新单元，使用输出信号值X和存储的阈值系数Deltatheta1和Deltatheta2来计算Deltatheta1X + Deltatheta2（X-1），并且通过将计算结果增加来更新存储在阈值存储单元中的阈值变量theta的值 。

公开(公告)号：US20100138043A1

公开(公告)日：2010-06-03

申请号：US12596461

申请日：2008-04-09

申请人： Hideki Kajima

发明人： Hideki Kajima

IPC分类号： B25J13/00

CPC分类号： B62D57/032

摘要： When a swinging leg (e.g., the leg link LR) lands on road surface, a control unit 14 included in a robot 100 controls an actuator 15 driving an ankle joint 122 to make the ankle joint 122 in a leg link LR soft and changes a real angle of the ankle joint 122 according to road profile, not to follow a prespecified trajectory of target angle. Further, after the leg link LR lands, the control unit 14 corrects the trajectory of the target angle of the ankle joint 122 to cancel out a difference between the real angle of the ankle joint 122 and the target angle. Moreover, the control unit 14 controls the actuator 15 to make the ankle joint 122 hard, so that the real angle of the ankle joint 122 of the leg link LR, a supporting leg, follows the corrected trajectory of the target angle.

摘要翻译： 当摆动腿（例如，腿部连杆LR）落在路面上时，包括在机器人100中的控制单元14控制致动器15驱动脚踝关节122，使腿部连杆LR中的踝关节122变软 踝关节122的实际角度根据道路轮廓，不遵循预定的目标角度轨迹。 此外，在腿连接LR着陆之后，控制单元14校正踝关节122的目标角度的轨迹，以抵消踝关节122的实际角度与目标角度之间的差异。 此外，控制单元14控制致动器15使踝关节122变硬，使得腿连杆LR的脚踝关节122的实际角度（支撑腿）跟随目标角度的校正轨迹。

公开(公告)号：US08554366B2

公开(公告)日：2013-10-08

申请号：US12596461

申请日：2008-04-09

申请人： Hideki Kajima

发明人： Hideki Kajima

IPC分类号： G06F19/00

CPC分类号： B62D57/032

摘要： When a swinging leg (e.g., the leg link LR) lands on road surface, a control unit 14 included in a robot 100 controls an actuator 15 driving an ankle joint 122 to make the ankle joint 122 in a leg link LR soft and changes a real angle of the ankle joint 122 according to road profile, not to follow a prespecified trajectory of target angle. Further, after the leg link LR lands, the control unit 14 corrects the trajectory of the target angle of the ankle joint 122 to cancel out a difference between the real angle of the ankle joint 122 and the target angle. Moreover, the control unit 14 controls the actuator 15 to make the ankle joint 122 hard, so that the real angle of the ankle joint 122 of the leg link LR, a supporting leg, follows the corrected trajectory of the target angle.

摘要翻译： 当摆动腿（例如，腿部连杆LR）落在路面上时，包括在机器人100中的控制单元14控制致动器15驱动脚踝关节122，使腿部连杆LR中的踝关节122变软 踝关节122的实际角度根据道路轮廓，不遵循预定的目标角度轨迹。 此外，在腿连接LR着陆之后，控制单元14校正踝关节122的目标角度的轨迹，以抵消踝关节122的实际角度与目标角度之间的差异。 此外，控制单元14控制致动器15使踝关节122变硬，使得腿连杆LR的脚踝关节122的实际角度（支撑腿）跟随目标角度的校正轨迹。

公开(公告)号：US08958976B2

公开(公告)日：2015-02-17

申请号：US12785123

申请日：2010-05-21

申请人： Hideki Kajima ,  Yoshiyuki Semba

发明人： Hideki Kajima ,  Yoshiyuki Semba

IPC分类号： G06F17/10 ,  A61G5/04 ,  A61G5/10 ,  B60L3/00 ,  B60L15/20

CPC分类号： A61G5/043 ,  A61G5/10 ,  A61G2203/42 ,  B60L3/0015 ,  B60L15/2009 ,  B60L2200/16 ,  B60L2200/34 ,  Y02T10/645 ,  Y02T10/7275

摘要： Provided is a traveling apparatus which performs desired traveling under an inverted pendulum control. The traveling apparatus includes: a ground contact member that operates to come into a ground-contact state or a non-ground-contact state with respect to a road surface; a drive portion that drives the ground contact member; a control portion that controls the drive portion; an abnormality degree calculation portion that calculates a degree of abnormality of a traveling state; and an upper bound calculation portion that changes an upper bound of a distance between the ground contact member and the road surface according to the degree of abnormality of the traveling state calculated by the abnormality degree calculation portion. The control portion controls driving of the drive portion so that the distance between the ground contact member and the road surface is smaller than or equal to the upper bound calculated by the upper bound calculation portion.

摘要翻译： 提供了在倒立摆控制下进行所需行驶的行驶装置。 行驶装置包括：地面接触构件，其相对于路面进行接地接触状态或非接地状态; 驱动所述接地构件的驱动部; 控制部，其控制所述驱动部; 异常度计算部，计算行驶状态的异常度; 以及根据由异常度计算部计算的行驶状态的异常程度，改变接地构件与路面之间的距离的上限的上限计算部。 控制部控制驱动部的驱动，使得接地部件与路面之间的距离小于或等于由上限计算部计算出的上限。

公开(公告)号：US08560152B2

公开(公告)日：2013-10-15

申请号：US12440980

申请日：2008-09-17

申请人： Hideki Kajima ,  Yoshiyuki Semba

发明人： Hideki Kajima ,  Yoshiyuki Semba

IPC分类号： G06F17/00 ,  G06F19/00 ,  B60T8/32 ,  B60T8/24

CPC分类号： G05D1/0891 ,  B62K11/007 ,  Y02T10/7258

摘要： To provide an inverted type moving body capable of continuing the inversion control with stability even when abnormality occurs in the inversion control while the moving body is moving, and a method of controlling the inverted type moving body. In an inverted type moving body including a rotational body having a circular cross section, a driving portion that rotationally drives the rotational body, a main body that supports the rotational body, and a control portion that maintains the inverted state of the main body by controlling the driving portion such that the rotational driving of the rotational body touching a floor surface is controlled, the control portion multiplies a signal obtained based on the inclined state of the main body by a predefined gain to calculate the driving amount of the rotational body for maintaining the inverted state, and reduces the gain when the inverted state of the main body is determined to be abnormal.

摘要翻译： 为了提供能够即使在移动体移动时的反转控制中发生异常的情况下也能够稳定地继续反转控制的反转型移动体，以及控制倒置型移动体的方法。 在包括具有圆形横截面的旋转体的倒置型移动体中，旋转驱动旋转体的驱动部，支撑旋转体的主体，以及通过控制来保持主体的倒转状态的控制部 所述驱动部使得所述旋转体接触地板表面的旋转驱动被控制，所述控制部将基于所述主体的倾斜状态获得的信号乘以预定的增益，以计算用于维持的所述旋转体的驱动量 反转状态，并且当主体的倒置状态被确定为异常时，降低增益。

公开(公告)号：US20100179749A1

公开(公告)日：2010-07-15

申请号：US12440980

申请日：2008-09-17

申请人： Hideki Kajima ,  Yoshiyuki Semba

发明人： Hideki Kajima ,  Yoshiyuki Semba

IPC分类号： G05D1/08 ,  B62K17/00

CPC分类号： G05D1/0891 ,  B62K11/007 ,  Y02T10/7258

摘要： To provide an inverted type moving body capable of continuing the inversion control with stability even when abnormality occurs in the inversion control while the moving body is moving, and a method of controlling the inverted type moving body. In an inverted type moving body including a rotational body having a circular cross section, a driving portion that rotationally drives the rotational body, a main body that supports the rotational body, and a control portion that maintains the inverted state of the main body by controlling the driving portion such that the rotational driving of the rotational body touching a floor surface is controlled, the control portion multiplies a signal obtained based on the inclined state of the main body by a predefined gain to calculate the driving amount of the rotational body for maintaining the inverted state, and reduces the gain when the inverted state of the main body is determined to be abnormal.

摘要翻译： 为了提供能够即使在移动体移动时的反转控制中发生异常的情况下也能够稳定地继续反转控制的反转型移动体，以及控制倒置型移动体的方法。 在包括具有圆形横截面的旋转体的倒置型移动体中，旋转驱动旋转体的驱动部，支撑旋转体的主体，以及通过控制来保持主体的倒转状态的控制部 所述驱动部使得所述旋转体接触地板表面的旋转驱动被控制，所述控制部将基于所述主体的倾斜状态获得的信号乘以预定的增益，以计算用于维持的所述旋转体的驱动量 反转状态，并且当主体的倒置状态被确定为异常时，降低增益。