US 20080042806 A1
A method of establishing wireless communications between an interrogator and individual ones of multiple wireless identification devices, the method comprising combining tree search and Aloha methods to establish communications between the interrogator and individual ones of the multiple wireless identification devices without collision. A system comprising an interrogator, and a plurality of wireless identification devices configured to communicate with the interrogator in a wireless fashion, the respective wireless identification devices having a unique identification number, the interrogator being configured to employ tree search and Aloha techniques to determine the unique identification numbers of the different wireless identification devices so as to be able to establish communications between the interrogator and individual ones of the multiple wireless identification devices without collision by multiple wireless identification devices attempting to respond to the interrogator at the same time.
41. A method for performing radio frequency communications, the method comprising:
(A) generating by a first radio frequency identification (RFID) tag a first random number;
(B) transmitting by an interrogator a first request for a first response from the first RFID tag if the first random number is within a first subgroup of possible random numbers;
(C) receiving by the first RFID tag the first request;
(D) communicating by the first RFID tag the first response when the first random number is within the first subgroup, the communicating being performed at a time based upon a second random number;
(E) redefining the first subgroup as a subset of a previous first subgroup and repeating steps B-E when a collision occurs; and
(F) redefining the first subgroup as another subgroup of possible random numbers when the first response is received by the interrogator without collision, the first response including one or more random numbers generated by the first RFID tag.
42. The method of
43. The method of
44. The method of
45. The method of
46. The method of
47. The method of
48. The method of
49. The method of
50. The method of
51. The method of
52. A method for a radio frequency identification (RFID) tag to perform wireless communications, the method comprising:
receiving a request, the request including a subset of possible random numbers;
determining whether a first random number generated by the RFID tag is within the subset; and
communicating a response to the request at a time based at least in part on a second random number generated by the RFID tag.
53. The method of
54. The method of
55. The method of
56. The method of
57. The method of
58. The method of
59. The method of
60. A method for an interrogator to poll a plurality of radio frequency identification (RFID) tags, the method comprising:
causing each of the plurality of RFID tags to generate one or more random numbers, each random number being generated by respective ones of the plurality of RFID tags independently of other ones of the plurality of RFID tags;
defining a first subgroup of a plurality of subgroups of possible random numbers;
transmitting a request to the plurality of RFID tags for identified RFID tags having generated a random number within the first subgroup to respond;
repeatedly redefining the first subgroup to include fewer possible random numbers and transmitting a new request for identified RFID tags having a random number within the first subgroup to respond until receiving a response without a collision or receiving no response; and
repeatedly picking a different subgroup and repeating the transmitting, the receiving, and the repeatedly redefining when a valid response or no response is received until the plurality of RFID tags have been identified, the valid response including at least one random number.
61. The method of
62. The method of
63. The method of
64. The method of
65. The method of
This invention relates to communications protocols and to digital data communications. Still more particularly, the invention relates to data communications protocols in mediums such as radio communication or the like. The invention also relates to radio frequency identification devices for inventory control, object monitoring, determining the existence, location or movement of objects, or for remote automated payment.
Communications protocols are used in various applications. For example, communications protocols can be used in electronic identification systems. As large numbers of objects are moved in inventory, product manufacturing, and merchandising operations, there is a continuous challenge to accurately monitor the location and flow of objects. Additionally, there is a continuing goal to interrogate the location of objects in an inexpensive and streamlined manner. One way of tracking objects is with an electronic identification system.
One presently available electronic identification system utilizes a magnetic coupling system. In some cases, an identification device may be provided with a unique identification code in order to distinguish between a number of different devices. Typically, the devices are entirely passive (have no power supply), which results in a small and portable package. However, such identification systems are only capable of operation over a relatively short range, limited by the size of a magnetic field used to supply power to the devices and to communicate with the devices.
Another wireless electronic identification system utilizes a large active transponder device affixed to an object to be monitored which receives a signal from an interrogator. The device receives the signal, then generates and transmits a responsive signal. The interrogation signal and the responsive signal are typically radio-frequency (RF) signals produced by an RF transmitter circuit. Because active devices have their own power sources, and do not need to be in close proximity to an interrogator or reader to receive power via magnetic coupling. Therefore, active transponder devices tend to be more suitable for applications requiring tracking of a tagged device that may not be in close proximity to an interrogator. For example, active transponder devices tend to be more suitable for inventory control or tracking.
Electronic identification systems can also be used for remote payment. For example, when a radio frequency identification device passes an interrogator at a toll booth, the toll booth can determine the identity of the radio frequency identification device, and thus of the owner of the device, and debit an account held by the owner for payment of toll or can receive a credit card number against which the toll can be charged. Similarly, remote payment is possible for a variety of other goods or services.
A communication system, such as a wireless identification system, typically includes two transponders: a commander station or interrogator, and a responder station or transponder device which replies to the interrogator.
If the interrogator has prior knowledge of the identification number of a device which the interrogator is looking for, it can specify that a response is requested only from the device with that identification number. Sometimes, such information is not available. For example, there are occasions where the interrogator is attempting to determine which of multiple devices are within communication range.
When the interrogator sends a message to a transponder device requesting a reply, there is a possibility that multiple transponder devices will attempt to respond simultaneously, causing a collision, and thus an erroneous message to be received by the interrogator. For example, if the interrogator sends out a command requesting that all devices within a communications range identify themselves, and gets a large number of simultaneous replies, the interrogator may not able to interpret any of these replies. Thus, arbitration schemes are employed to permit communications free of collisions.
In one arbitration scheme or system, described in commonly assigned U.S. Pat. Nos. 5,627,544; 5,583,850; 5,500,650; and 5,365,551, all to Snodgrass et al. and all incorporated herein by reference, the interrogator sends a command causing each device of a potentially large number of responding devices to select a random number from a known range and use it as that device's arbitration number. By transmitting requests for identification to various subsets of the full range of arbitration numbers, and checking for an error-free response, the interrogator determines the arbitration number of every responder station capable of communicating at the same time. Therefore, the interrogator is able to conduct subsequent uninterrupted communication with devices, one at a time, by addressing only one device.
Another arbitration scheme is referred to as the Aloha or slotted Aloha scheme. This scheme is discussed in various references relating to communications, such as Digital Communications: Fundamentals and Applications, Bernard Sklar, published January 1988 by Prentice Hall. In this type of scheme, a device will respond to an interrogator using one of many time domain slots selected randomly by the device. A problem with the Aloha scheme is that if there are many devices, or potentially many devices in the field (i.e. in communications range, capable of responding) then there must be many available slots or many collisions will occur. Having many available slots slows down replies. If the magnitude of the number of devices in a field is unknown, then many slots are needed. This results in the system slowing down significantly because the reply time equals the number of slots multiplied by the time period required for one reply.
An electronic identification system which can be used as a radio frequency identification device, arbitration schemes, and various applications for such devices are described in detail in commonly assigned U.S. patent application Ser. No. 08/705,043, filed Aug. 29, 1996, and incorporated herein by reference.
The invention provides a wireless identification device 1, configured to provide a signal to identify the device in response to an interrogation signal.
One aspect of the invention provides a method of establishing wireless communications between an interrogator and individual ones of multiple wireless identification devices. Tree search and Aloha methods are combined to establish communications between the interrogator and individual ones of the multiple wireless identification devices without collision.
One aspect of the invention provides a method of addressing messages from an interrogator to a selected one or more of a number of communications devices. A first predetermined number of bits are established to be used as unique identification numbers. Unique identification numbers respectively having the first predetermined number of bits are established for respective devices. A second predetermined number of bits are established to be used for random values. The devices are caused to select random values. Respective devices choose random values independently of random values selected by the other devices. The interrogator transmits a command requesting devices having random values within a specified group of random values to respond, the specified group being less than or equal to the entire set of random values. Devices receiving the command respectively determine if their chosen random values fall within the specified group and, if so, send a reply to the interrogator within a randomly selected time slot of a number of slots. If not, they do not send a reply. The interrogator determines if a collision occurred between devices that sent a reply and, if so, creates a new, smaller, specified group.
One aspect of the invention provides a communications system comprising an interrogator, and a plurality of wireless identification devices configured to communicate with the interrogator in a wireless fashion. The respective wireless identification devices have a unique identification number. The interrogator is configured to employ tree search and Aloha techniques to determine the unique identification numbers of the different wireless identification devices so as to be able to establish communications between the interrogator and individual ones of the multiple wireless identification devices without collision by multiple wireless identification devices attempting to respond to the interrogator at the same time.
Another aspect of the invention provides a system comprising an interrogator configured to communicate to a selected one or more of a number of communications devices, and a plurality of communications devices. The devices are configured to select random values. Respective devices choose random values independently of random values selected by the other devices. The interrogator is configured to transmit a command requesting devices having random values within a specified group of random values to respond, the specified group being less than or equal to the entire set of random values. Devices receiving the command are configured to respectively determine if their chosen random values fall within the specified group and, if so, send a reply to the interrogator within a randomly selected time slot of a number of slots. If not, they do not send a reply. The interrogator is configured to determine if a collision occurred between devices that sent a reply and, if so, create a new, smaller, specified group.
One aspect of the invention provides a radio frequency identification device comprising an integrated circuit including a receiver, a transmitter, and a microprocessor. In one embodiment, the integrated circuit is a monolithic single die single metal layer integrated circuit including the receiver, the transmitter, and the microprocessor. The device of this embodiment includes an active transponder, instead of a transponder which relies on magnetic coupling for power, and therefore has a much greater range.
Preferred embodiments of the invention are described below with reference to the following accompanying drawings.
This disclosure of the invention is submitted in furtherance of the constitutional purposes of the U.S. Patent Laws “to promote the progress of science and useful arts” (Article 1, Section 8).
The device 12 transmits and receives radio frequency communications to and from an interrogator 26. An exemplary interrogator is described in U.S. patent application Ser. No. 08/907,689, filed Aug. 8, 1997 and incorporated herein by reference. Preferably, the interrogator 26 includes an antenna 28, as well as dedicated transmitting and receiving circuitry, similar to that implemented on the integrated circuit 16.
Generally, the interrogator 26 transmits an interrogation signal or command 27 via the antenna 28. The device 12 receives the incoming interrogation signal via its antenna 14. Upon receiving the signal 27, the device 12 responds by generating and transmitting a responsive signal or reply 29. The responsive signal 29 typically includes information that uniquely identifies, or labels the particular device 12 that is transmitting, so as to identify any object or person with which the device 12 is associated.
Although only one device 12 is shown in
The radio frequency data communication device 12 can be included in any appropriate housing or packaging. Various methods of manufacturing housings are described in commonly assigned U.S. patent application Ser. No. 08/800,037, filed Feb. 13, 1997, and incorporated herein by reference.
If the power source 18 is a battery, the battery can take any suitable form. Preferably, the battery type will be selected depending on weight, size, and life requirements for a particular application. In one embodiment, the battery 18 is a thin profile or button-type cell forming a small, thin energy cell more commonly utilized in watches and small electronic devices requiring a thin profile. A conventional cell has a pair of electrodes, an anode formed by one face and a cathode formed by an opposite face. In an alternative embodiment, the power source 18 comprises to a series connected pair of cells. Instead of using a battery, any suitable power source can be employed.
The circuitry 16 further includes a backscatter transmitter and is configured to provide a responsive signal to the interrogator 26 by radio frequency. More particularly, the circuitry 16 includes a transmitter, a receiver, and memory such as is described in U.S. patent application Ser. No. 08/705,043.
Radio frequency identification has emerged as a viable and affordable alternative to tagging or labeling small to large quantities of items. The interrogator 26 communicates with the devices 12 via an RF link, so all transmissions by the interrogator 26 are heard simultaneously by all devices 12 within range.
If the interrogator 26 sends out a command requesting that all devices 12 within range identify themselves, and gets a large number of simultaneous replies, the interrogator 26 may not be able to interpret any of these replies. Therefore, arbitration schemes are provided.
If the interrogator 26 has prior knowledge of the a identification number of a device 12 which the interrogator 26 is looking for, it can specify that a response is requested only from the device 12 with that identification number. To target a command at a specific device 12, (i.e., to initiate point-on-point communication), the interrogator 26 must send a number identifying a specific device 12 along with the command. At start-up, or in so a new or changing environment, these identification numbers are not known by the interrogator 26. Therefore, the interrogator 26 must identify all devices 12 in the field (within communication range) such as by determining the identification numbers of the devices 12 in the field. After this is accomplished, point-to-point communication can proceed as desired by the interrogator 26.
Generally speaking, RFID systems are a type of multiaccess communication system. The distance between the interrogator 26 and devices 12 within the field is typically fairly short (e.g., several meters), so packet transmission time is determined primarily by packet size and baud rate. Propagation delays are negligible. In RFID systems, there is a potential for a large number of transmitting devices 12 and there is a need for the interrogator 26 to work in a changing environment, where different devices 12 are swapped in and out frequently (e.g., as inventory is added or removed). The inventors have determined that, in such systems, the use of random access methods work effectively for contention resolution (i.e., for dealing with collisions between devices 12 attempting to respond to the interrogator 26 at the same time).
RFID systems have some characteristics that are different from other communications systems. For example, one characteristic of the illustrated RFID systems is that the devices 12 never communicate without being prompted by the interrogator 26. This is in contrast to typical multiaccess systems where the transmitting units operate more independently. In addition, contention for the communication medium is short lived as compared to the ongoing nature of the problem in other multiaccess systems. For example, in a RFID system, after the devices 12 have been identified, the interrogator can communicate with them in a point-to-point fashion. Thus, arbitration in a RFID system is a transient rather than steady-state phenomenon. Further, the capability of a device 12 is limited by practical restrictions on size, power, and cost. The lifetime of a device 12 can often be measured in terms of number of transmissions before battery power is lost. Therefore, one of the most important measures of system performance in RFID arbitration is total time required to arbitrate a set of devices 12. Another measure is power consumed by the devices 12 during the process. This is in contrast to the measures of throughput and packet delay in other types of multiaccess systems.
Three variables are used: an arbitration value (AVALUE), an arbitration mask (AMASK), and a random value ID (RV). The interrogator sends a command causing each device of a potentially large number of responding devices to select a random number from a known range and use it as that device's arbitration number. The interrogator sends an arbitration value (AVALUE) and an arbitration mask (AMASK) to a set of devices 12. The receiving devices 12 evaluate the following equation:
(AMASK & AVALUE)==(AMASK & RV) wherein “&” is a bitwise AND function, and wherein “==” is an equality function. If the equation evaluates to “1,” (TRUE), then the device 12 will reply. If the equation evaluates to “0” (FALSE), then the device 12 will not reply. By performing this in a structured manner, with the number of bits in the arbitration mask being increased by one each time, eventually a device 12 will respond with no collisions. Thus, a binary search tree methodology is employed.
An example using actual numbers will now be provided using only four bits, for simplicity, reference being made to
Assume, for this example, that there are two devices 12 in the field, one with a random value (RV) of 1100 (binary), and another with a random value (RV) of 1010 (binary). The interrogator is trying to establish communications without collisions being caused by the two devices 12 attempting to communicate at the same time.
The interrogator sets AVALUE to 0000 (or all “don't care”, indicated by the character “X” in
Next, the interrogator sets AMASK to 0001 and AVALUE to 0000 and transmits an identify command. Both devices 12 in the field have a zero for their least significant bit, and (AMASK & AVALUE)==(AMASK & RV) will be true for both devices 12. For the device 12 with a random value of 1100, the left side of the equation is evaluated as follows (0001 & 0000)=0000. The right side is evaluated as (0001 & 1100)=0000. The left side equals the right side, so the equation is true for the device 12 with the random value of 1100. For the device 12 with a random value of 1010, the left side of the equation is evaluated as (0001 & 0000)=0000. The right side is evaluated as (0001 & 1010)=0000. The left side equals the right side, so the equation is true for the device 12 with the random value of 1010. Because the equation is true for both devices 12 in the field, both devices 12 in the field respond, and there is another collision.
Recursively, the interrogator next sets AMASK to 0011 with AVALUE still at 0000 and transmits an identify command. (AMASK & AVALUE)-==(AMASK & RV) is evaluated for both devices 12. For the device 12 with a random value of 1100, the left side of the equation is evaluated as follows (0011 & 0000)=0000. The right side is evaluated as (0011 & 1100)=0000. The left side equals the right side, so the equation is true for the device 12 with the random value of 1100, so this device 12 responds. For the device 12 with a random value of 1010, the left side of the equation is evaluated as (0011 & 0000)=0000. The right side is evaluated as (0011 & 1010)=0010. The left side does not equal the right side, so the equation is false for the device 12 with the random value of 1010, and this device 12 does not respond. Therefore, there is no collision, and the interrogator can determine the identity (e.g., an identification number) for the device 12 that does respond.
De-recursion takes place, and the devices 12 to the right for the same AMASK level are accessed by setting AVALUE at 0010 and using the same AMASK value 0011.
The device 12 with the random value of 1010 receives a command and evaluates the equation (AMASK & AVALUE)-==(AMASK & RV). The left side of the equation is evaluated as (0011 & 0010)=0010. The right side of the equation is evaluated as (0011 & 1010)=0010. The right side equals the left side, so the equation is true for the device 12 with the random value of 1010. Because there are no other devices 12 in the subtree, a good reply is returned by the device 12 with the random value of 1010. There is no collision, and the interrogator can determine the identity (e.g., an identification number) for the device 12 that does respond.
By recursion, what is meant is that a function makes a call to itself. In other words, the function calls itself within the body of the function. After the called function returns, de-recursion takes place and execution continues at the place just after the function call; i.e. at the beginning of the statement after the function call.
For instance, consider a function that has four statements (numbered 1, 2, 3, 4) in it, and the second statement is a recursive call. Assume that the fourth statement is a return statement. The first time through the loop (iteration 1) the function executes the statement 2 and (because it is a recursive call) calls itself causing iteration 2 to occur. When iteration 2 gets to statement 2, it calls itself making iteration 3. During execution in iteration 3 of statement 1, assume that the function does a return. The information that was saved on the stack from iteration 2 is loaded and the function resumes execution at statement 3 (in iteration 2), followed by the execution of statement 4 which is also a return statement. Since there are no more statements in the function, the function de-recurses to iteration 1. Iteration 1, had previously recursively called itself in statement 2. Therefore, it now executes statement 3 (in iteration 1). Following that it executes a return at statement 4. Recursion is known in the art.
Consider the following code, which employs recursion, and which can be used to implement operation of the method shown in
The symbol “<<” a represents a bitwise left shift. “<<1” means shift left by one place. Thus, 0001<<1 would be 0010. Note, however, that AMASK is originally called with a value of zero, and 0000<<1 is still 0000. Therefore, for the first recursive call, AMASK=(AMASK<<1)+1. So for the first recursive call, the value of AMASK is 0000+0001=0001. For the second call, AMASK=(0001<<1)+1=0010+1=0011. For the third recursive call, AMASK=(0011<<1)+1=0110+1=0111.
The routine generates values for AMASK and AVALUE to be used by the interrogator in an identify command “IdentifyCmnd.” Note that the routine calls itself if there is a collision. De-recursion occurs when there is no collision. AVALUE and AMASK would have values such as the following assuming there are collisions all the way down to the bottom of the tree.
This sequence of AMASK, AVALUE binary numbers assumes that there are collisions all the way down to the bottom of the tree, at which point the Identify command sent by the interrogator is finally successful so that no collision occurs. Rows in the table for which the interrogator is successful in receiving a reply without collision are marked with the symbol “*”. Note that if the Identify command was successful at, for example, the third line in the table then the interrogator would stop going down that branch of the tree and start down another, so the sequence would be as shown in the following table.
This method is referred to as a splitting method. It works by splitting groups of colliding devices 12 into subsets that are resolved in turn. The splitting method can also be viewed as a type of tree search. Each split moves the method one level deeper in the tree. Either depth-first or breadth first traversals of the tree can be employed.
Another arbitration method that can be employed is referred to as the “Aloha” method. In the Aloha method, every time a device 12 is involved in a collision, it waits a random period of time before retransmitting. This method can be improved by dividing time into equally sized slots and forcing transmissions to be aligned with one of these slots. This is referred to as “slotted Aloha.” In operation, the interrogator asks all devices 12 in the field to transmit their identification numbers in the next time slot. If the response is garbled, the interrogator informs the devices 12 that a collision has occurred, and the slotted Aloha scheme is put into action. This means that each device 12 in the field responds within an arbitrary slot determined by a randomly selected value. In other words, in each successive time slot, the devices 12 decide 14 to transmit their identification number with a certain probability.
The Aloha method is based on a system operated by the University of Hawaii. In 1971, the University of Hawaii began operation of a system named Aloha. A communication satellite was used to interconnect several university computers by use of a random access protocol. The system operates as follows. Users or devices transmit at any time they desire. After transmitting, a user listens for an acknowledgment from the receiver or interrogator. Transmissions from different users will sometimes overlap in time (collide), causing reception errors in the data in each of the contending messages. The errors are detected by the receiver, and the receiver sends a negative acknowledgment to the users. When a negative acknowledgment is received, the messages are retransmitted by the colliding users after a random delay. If the colliding users attempted to retransmit without the random delay, they would collide again. If the user does not receive either an acknowledgment or a negative acknowledgment within a certain amount of time, the user “times out” and retransmits the message.
In the slotted Aloha scheme, a sequence of coordination pulses is broadcast to all stations (devices). As is the case with the pure Aloha scheme, packet lengths are constant. Messages are required to be sent in a slot time between synchronization pulses, and can be started only at the beginning of a time slot. This reduces the rate of collisions because only messages transmitted in the same slot can interfere with one another. The retransmission mode of the pure Aloha scheme is modified for slotted Aloha such that if a negative acknowledgment occurs, the device retransmits after a random delay of an integer number of slot times.
Another form of Aloha scheme is called reservation-Aloha. The reservation-Aloha system has two basic modes: an unreserved mode, and a reserved mode.
In the unreserved mode, a time frame is established and divided into a number of small reservation subslots. Users (devices) use these subslots to reserve message slots. After requesting a reservation, the user (device) listens for an acknowledgment and a slot assignment.
In the reserved mode, a time frame is divided into a certain number of slots whenever a reservation is made. All but the last slot are used for message transmissions. The last slot is subdivided into subslots to be used for reservations. Users (devices) send message packets in their assigned portions of the slots reserved for message transmissions.
In one embodiment, the same randomly selected time slot is used by a device 12 at different levels of the tree (i.e., for different values of AMASK and AVALUE). In another embodiment, different randomly selected times slots are used by a device 12 at different levels of the tree (i.e., for different values of AMASK and AVALUE). In one embodiment, a combination of these approaches is used. For example, one embodiment utilizes a method where the interrogator goes down the tree until some responses without collision are received, before the devices 12 re-randomize their Aloha random number. This can be classified as an adaptive method. Other adaptive methods are possible. For example, in one embodiment, the number of Aloha slots is reduced at lower levels of the tree. The number of slots can be reduced by the same number for each level down the tree, or by a number that varies depending on the number of levels down the tree. Thus, for example, the number of slots can remain constant through a progression down the tree until some responses without collision are received, at which point the number of slots is reduced.
Thus, this embodiment provides the advantages of both the Aloha methods and the tree sorting methods of establishing communications without collisions.
In another embodiment, levels of the search tree are skipped. Skipping levels in the tree, after a collision caused by multiple devices 12 responding, reduces the number of subsequent collisions without adding significantly to the number of no replies. In real-time systems, it is desirable to have quick arbitration sessions on a set of devices 12 whose unique identification numbers are unknown. Level skipping reduces the number of collisions, both reducing arbitration time and conserving battery life on a set of devices 12. In one embodiment, every other level is skipped. In alternative embodiments, more than one level is skipped each time.
The trade off that must be considered in determining how many (if any) levels to skip with each decent down the tree is as follows. Skipping levels reduces the number of collisions, thus saving battery power in the devices 12. Skipping deeper (skipping more than one level) further reduces the number of collisions. The more levels that are skipped, the greater the reduction in collisions. However, skipping levels results in longer search times because the number of queries (Identify commands) increases. The more levels that are skipped, the longer the search times. Skipping just one level has an almost negligible effect on search time, but drastically reduces the number of collisions. If more than one level is skipped, search time increases substantially. Skipping every other level drastically reduces the number of collisions and saves battery power without significantly increasing the number of queries.
Level skipping methods are described in a commonly assigned patent application (attorney docket MI40-117) naming Clifton W. Wood, Jr. and Don Hush as inventors, titled “Method of Addressing Messages, Method of Establishing Wireless Communications, and Communications System,” filed concurrently herewith, and incorporated herein by reference.
In compliance with the statute, the invention has been described in language more or less specific as to structural and methodical features. It is to be understood, however, that the invention is not limited to the specific features shown and described, since the means herein disclosed comprise preferred forms of putting the invention into effect. The invention is, therefore, claimed in any of its forms or modifications within the proper scope of the appended claims appropriately interpreted in accordance with the doctrine of equivalents.